id
stringlengths 24
24
| idx
int64 0
402
| paragraph
stringlengths 106
17.2k
|
|---|---|---|
66eb9a1712ff75c3a1c5a2be | 7 | Another important property of cg-2 is its ability to maintain spin state, which is crucial for exploring spin-based applications . In radical systems, spin relaxation from excited state to ground state is mainly triggered by the interactions with surrounding environment or itself, as the rate of spontaneous relaxation is negligible. Thereby, the spin-lattice relaxation time (T1) and coherence time (Tm) of cg-2 were measured to be 96 μs and 0.8 μs at 100 K and the field of 3401 G (Supplementary Fig. ), respectively, using inversion-recovery method (π-T-π/2-τ-π-τ-echo) and modified Hahn-echo pulse sequency (π/2-τ-π-τ-echo). The relatively shorter relaxation time of cg-2, compared with typical carbon-centered radicals , are probably due to the influence of numerous magnetic proton nuclei on the substituents. Moreover, the Rabi frequency 𝛺𝛺 𝑚𝑚 ±1 ↔𝑚𝑚 0 of 8.21 MHz for cg-2 provides the experimental timescale (τspin-flip = 61 ns) for one spin-flip operation |
66eb9a1712ff75c3a1c5a2be | 8 | Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) measurements of cg-2 in dry CH2Cl2 revealed two successive redox waves, Ⅰ and Ⅱ with the half-wave potentials E1/2 at -0.21 V and 0 V (all values vs Fc/Fc + ), respectively (Fig. ). The redox waves Ⅰ and Ⅱ correspond to stepwise one-electron oxidations of cg-2 from its neutral diradical state to the radical cation state, then the dication state. These oxidation states were obtained by chemical oxidation with NO•SbF6 and characterized by UV-vis-NIR absorption spectroscopy (Fig. ). The neutral diradical state of cg-2 shows weak absorption bands at 675 nm and longer wavelengths region. According to timedependent DFT calculation, these weak absorption bands could be assigned to the forbidden transitions from SOMO-α to LUMO-α and SOMO-β to LUMO-β (Supplementary Fig. ). The radical cation and dication states of cg-2 exhibit similar absorption features: an intense band at 669 nm and a broad band centered at 1022 and 924 nm, respectively, indicating the structural similarity in their molecular orbitals related to electronic transitions (Supplementary Fig. ). |
66eb9a1712ff75c3a1c5a2be | 9 | Similar to the neutral diradical state, the electron and charge in the radical cation state of cg-2 are also spatially segregated (i.e., the electron is distributed on one half, and the positive charge is distributed on the other half). This unique electronic property of the radical cation state of cg-2 is also predicted by DFT calculated spin density distribution and electrostatic potential surface (Fig. ). cw-EPR spectrum of the radical cation state of cg-2 was measured in frozen CH2Cl2 at 153 K to suppress the rapid tumbling motion in solution. The signal is split into a quintet by four anisotropic protons with simulated hyperfine coupling constants of aH = 0.40, 0.39, 0.43 and 0.52 mT (Fig. ). This observation provides direct experimental evidence for the spatially segregated spin in the radical cation state of cg-2, because if the spin is globally delocalized over the whole molecule, the EPR signal will be split into a nonet by eight anisotropic protons, each with the halved aH. It should be noted that as the temperature increases, a nonet EPR signal appears, probably due to the averaging effect caused by tumbling motion or by thermally populated electron hoping from one half to the other half (Supplementary Fig. ). The dication state of cg-2 recovers its closed-shell electronic structure, therefore the 1 H NMR measurement was performed. The protons Ha and Hb appear at downfield region (Fig. ), indicating the deshielding effect brought by the presence of positive charge. |
66eb9a1712ff75c3a1c5a2be | 10 | In summary, we realized the solution-phase synthesis and isolation of the Clar's goblet derivative cg-2 with bulky substituents. The precise construction of its topologically frustrated sp 2 -C network was achieved through an intermolecular radical-radical coupling approach. The synthetic scalability and kinetical inertness of cg-2 allow us to experimentally elucidate the long-standing doubt about the interaction between the two spins in Clar's goblet. Magnetic studies reveal that the two spins are spatially segregated in Clar's goblet with an average distance of 8.7 Å, and are AFM coupled in the ground state. The singlet ground state can be thermally excited to a triplet state with an ∆ES-T of -0.29 kcal/mol. Furthermore, the spatial segregation of spin/charge can be directly observed through EPR splitting in the radical cation state of cg-2. This spin pairing discovered in Clar's goblet not only suggests a spin entanglement phenomenon at the molecular scale, but also may enable practical logic operations in spin-based quantum computation. Finally, we hope this study could inspire and motivate further efforts to uncover the fundamental mechanism of this spin entanglement and push the application of tailor-made molecular spins in quantum information technologies. |
64c3a8d8658ec5f7e5517a32 | 0 | Photochemical reactions mediated by tetrabutylammonium decatungstate (TBADT), a versatile and inexpensive photocatalyst, have been extensively studied over the past decade. The mainstream use of the decatungstate photocatalysis relies on its property to easily abstract H-atoms (hydrogen atom transfer, HAT) from C-H bonds with bond dissociation energies of up to 100 kcal mol -1 upon photoexcitation with UV light (310-400 nm, Scheme 1, a). This keynote property has resulted in the emergence of several facile C-H activation methodologies, such as the works of Fagnoni, MacMillan, Noёl, and Melchiorre groups, to name a few. On the contrary, even though the decatungstate anion is a strong one-electron oxidant in its photoexcited state (E1/2 *[W10O32] 4-/ [W10O32] 5- = +2.4 V vs SCE), the applications of decatungstate as a photoredox catalyst remain largely underdeveloped (Scheme 1, b). To the best of our knowledge, such transformations have been limited to SET oxidation of aromatic hydrocarbons and to a few synthetic applications reported by Fagnoni and Ravelli (Scheme 1, c), who developed benzylation of electrophilic alkenes via TBADT-photocatalyzed generation of resonance-stabilized benzylic radicals by oxidation of benzylsilanes and arylacetic acids. In view of our expertise in the cyclopropanol chemistry, we envisioned that SET oxidation of readily available tertiary cyclopropanols 1 by the photoexcited decatungstate should be a feasible process (Ep of cyclopropanols = +1.0 -+2.0 V vs SCE) while HAT and homolytic cleavage of the O-H bond are unlikely to occur due to the high dissociation energy of the O-H bond in aliphatic alcohols (typical BDE values 103-105 kcal mol ). The single electron oxidation of 1 would result in the generation of radical cation A, which readily undergoes cyclopropane ring opening to afford the β-ketoalkyl radical B. Although B is an unstablilized primary alkyl radical, the process is driven by the release of cyclopropane ring strain (Scheme 1, d). Similarly to the previous works, we expected that the produced radicals B could be intercepted by electron-deficient alkenes 2 to furnish a new C-C bond in the respective adducts 3. Although β-ketoalkyl radical-mediated transformations are widespread in the chemistry of cyclopropanols, not much attention has been paid to generate these species in a photocatalytic manner, and especially to enable their reactions with electron-deficient alkenes. The group of Yamamoto used proton-coupled electron transfer (PCET) strategy for the intramolecular reaction of cyclopropanols and electron-deficient alkenes leading to the ring expansion. Hu and co-workers developed chlorine radical-induced reaction of alcohols, including cyclopropanols, and alkenes catalyzed by iron salts under blue LED irradiation. In the work of Melchiorre et al., enantioselective photochemical cascade reaction of cyclopropanols and α,βunsaturated aldehydes was developed, which combined the excited-state and ground-state reactivity of chiral organocatalytic intermediates. Traditionally, in similar non-photochemical reactions, transition metal derivatives (Ag(I), Fe(III), and Mn(III) ) have been applied as catalysts, often in combination with additional oxidative reagents and under harsh reaction conditions. Moreover, stoichiometric oxidants can be used on their own, without any catalytic assistance. Here we present a new photocatalytic methodology that enables redox-neutral addition of cyclopropanolderived β-ketoalkyl radicals B to electrophilic olefins. The methodology utilizes TBADT as a readily available photoredox catalyst, features a broad substrate scope, and operates under ambient temperature conditions. Notably, no competitive HAT processes have been observed and sensitive C-H bonds remained intact in the obtained products 3. Mechanistic studies have confirmed the role of photoexcited decatungstate as a SET oxidant. Moreover, we have revealed the existence of photoactive electron donoracceptor (EDA) complexes between aromatic cyclopropanols 1 and electron-deficient alkenes 2. The EDA complexes can produce the corresponding adducts 3 upon photoexcitation, even without the assistance of the photoredox catalyst. |
64c3a8d8658ec5f7e5517a32 | 1 | Optimization studies were performed on the model cross-coupling reaction of p-methoxyphenyl cyclopropanol 1a with electrophilic benzylidene malononitrile 2a under ultraviolet light (370 nm) irradiation (Table ; see also Table in the Supporting Information for additional data). The experiments were carried out in a 3D printed reactor designed by Noёl group and equipped with a Kessil lamp (43 W at 100% light intensity). Previously reported conditions (2 mol% TBADT, acetonitrile) were employed in the initial experiment (entry 1) which delivered ketone 3a in 84% yield after 3 h of irradiation. However, slightly improved 88-90% yields and shorter reaction time (1 h) were achieved at lower catalyst loading (1 mol%) when acetonitrile was replaced with dichloromethane (CH2Cl2) or acetonitrile/water mixture (entry 2 and 3). To confirm the photochemical nature of the process, a control experiment without light was carried out and no reaction occurred (entry 4). Moreover, the reaction was sensitive to the light intensity, as the conversion decreased when the lamp was used at lower power output (see Table in the Supporting Information). Surprisingly, 3a was obtained in 25% yield even in the absence of the photocatalyst (entry 5) indicating a plausible slower reaction via EDA complex of 1a and 2a (see discussion below). However, TBADT was crucial to accelerate the reaction and benzophenone (10 mol%) as an alternative photocatalyst afforded 3a in only 44% yield (entry 6). Table . Optimization of the reaction conditions. Entry Deviations Yield, % With the optimal reaction conditions established (CH2Cl2 as a solvent, 1 mol% TBADT), we explored the scope of different cyclopropanols in their photocatalytic cross-coupling reactions with benzylidene malononitrile 2a (Scheme 2, a). Cyclopropanols with electron-rich aromatic substituents afforded β-functionalized ketones in 56-87% yields (compounds 3aa-3ia). Among them, cyclopropanols 1b and 1c with MeO-substituents in meta-and ortho-positions gave the respective products 3ba and 3ca in lower 61% and 56% yields compared to para-methoxyphenyl cyclopropanol (3aa, 87% yield), which could be caused by steric factors or somewhat different electronic properties. For the reactions yielding m-OMesubstituted 3ba and 2-naphthyl derivative 3da, acetonitrile/water mixture was more beneficial reaction media compared to CH2Cl2, in which lower yields were obtained (e.g., 71% vs 59% for 3da). Notably, pyrrolyl and thiophenyl substituents tolerated the reaction conditions and the corresponding products 3ha and 3ia were obtained in 57% and 88% yields, respectively. The reaction with cyclopropanol 1j, bearing an electron-withdrawing p-trifluoromethyl group, afforded the product 3ja in a moderate 54% yield. However, less electron-deficient p-chlorophenyl cyclopropanol 1k was a successful substrate, which produced the corresponding adduct 3ka in excellent 92% yield. By contrast to aryl-substituted cyclopropanols 1a-k, alkyl and alkenyl cyclopropanols 1l-p and were less reactive and afforded the respective products 3la-3pa in low to moderate yields (42-61%) after significantly longer reaction time (20-48 h; Scheme 2, b). Thus, in the reactions of alkenyl cyclopropanols, 3na was produced from 1-cyclohexenyl cyclopropanol 1n in a low 41% yield. Styryl cyclopropanol 1o behaved similarly resulting in the formation of 3oa in 42% yield. N-Boc-protected cyclopropanol 1p was a competent substrate that produced 3pa in 40% yield, while its N-benzyl analogue 1r was incompatible and resulted in a complex mixture in which no respective cross-coupling product was detected. Additionally, we found that oxidants such as N-fluorobenzenesulfonimide (NFSI) or potassium persulfate as substoichiometric additives (0.2-0.5 equiv.) can improve yields, accelerate, or even trigger the reaction of otherwise unreactive cyclopropanols. For instance, compound 3la was obtained in 63% yield after 3 h with NFSI additive (0.5 equiv.) while the standard conditions delivered only 53% yield after 20 h. Similarly, n-amylcyclopropanol 1q was converted into the corresponding ketone product 3qa. Besides highly electrophilic alkene 2a, 2-phenylethyl cyclopropanol 1l was also able to react with dimethyl fumarate 2f and diethyl benzylidenemalonate 2k, affording the respective products 3lf and 3lk in 58% and 40% yields, while only traces of the products were formed without NFSI additive. |
64c3a8d8658ec5f7e5517a32 | 2 | Various benzylidene malononitriles with electron-donating and -withdrawing substituents in the phenyl moiety provided the respective products in 68-86% yields (compounds 3ab, 3ac and 3ad). However, complex reaction mixture and negligible (5-12%) yield of the respective adducts 3 were observed in the reactions with heteroarylidene malononitriles 2n and 2o, as well as with a substrate with a single nitrile group 2p. In addition, no product was found in the reaction of cyclopropanol 1a with methyl-substituted benzylidene malononitrile 2q, which was unreactive under the standard conditions and in the presence of NFSI, probably due to steric hindrance caused by a methyl group. Fumaronitrile 2e, dimethyl fumarate 2f and vinyl phenyl sulfone 2g were alkylated efficiently furnishing the respective products 3ae-ag in 63-85% yields. Notably, products synthesized from alkylidene oxindoles were isolated in good yields (3ah and 3ai, 65% and 66% yield, respectively). Compounds containing an oxindole moiety in their structure have demonstrated a wide range of biological activities. Phenylmaleimide 2j was converted to the corresponding ketone 3aj in good 71% yield. However, similar maleic anhydride 2r was unreactive and only decomposition products of cyclopropanol 1a were observed. Malonate and 1,3-diketone derivatives underwent the desired transformation more sluggish than the rest of the substrates, nevertheless, compounds 3ak, 3al and 3am were isolated in acceptable 44-50% yields. The reaction with less electrondeficient triphenylethylene 2s failed to deliver the corresponding adduct. The evident influence of the electrophilic properties of the alkene indicated that it was not an insignificant bystander in the reaction, merely intercepting the cyclopropanol-derived β-ketoalkyl radical. Instead, it plays a crucial role in the reaction mechanism (see discussion below). |
64c3a8d8658ec5f7e5517a32 | 3 | The batch reactions of cyclopropanols 1a and 1l with benzylidene malononitrile 2a have been successfully upscaled when translated into a continuous-flow process (Figure , a). Flow chemistry is especially advantageous for performing photochemical transformations, since it ensures more efficient light irradiation, consistency of results, and scalability among other benefits. Recently, TBADT recycling strategy by nanofiltration was developed by the group of Noёl, allowing to use TBADT-photocatalyzed reactions at industrial scale. After a brief optimization (see Tables and in the Supporting Information), we were able to prepare compounds 3aa and 3la in good 74-75% yields and gram amounts with 20-30 min of residence time. |
64c3a8d8658ec5f7e5517a32 | 4 | To demonstrate the synthetic utility of the obtained products containing multiple activated sites with prominent and diverse reactivity, compound 3n was converted into a variety of derivatives (Figure ). First, we aimed to elucidate the nature of minor (ca. 10% yield) by-products that emerged during the purification of 3la and similar adducts of benzylidene malononitrile by column chromatography on silica gel. Based on literature precedents, we surmised that the by-products might result from a silicacatalyzed intramolecular aldol-type cyclization. Taking advantage of the observed transformation, we subjected compound 3la to the reaction with trimethylsilyl triflate (TMSOTf) and triethylamine which resulted in TMS-protected cyclopentanol 4a with excellent 95% yield. Likewise, the analogous reaction of 3aa was equally successful and implies that such cyclization could also be performed with other similar cross-coupling products 3. Besides the cyclization reaction, the rich chemistry of malononitrile and carbonyl moieties in 3la also allowed its facile conversion into methyl ester 4b (85% yield) by oxidation with molecular oxygen, synthesis of amide 4c in 65% yield by Cu-catalyzed hydrolysis of a nitrile group, and featured alcohol 4d by reduction with NaBH4. The investigation of the substrate scope revealed that the electrophilic properties of alkenes are crucial for the reaction. Moreover, benzylidene malononitrile 2a, which is quite strong electron acceptor and oneelectron oxidant, could react with cyclopropanol 1a bearing an electron-rich arene moiety even without the catalytic assistance of TBADT catalyst, though the reaction is slower (Table , entry 1 vs 5). This observation can be explained by the formation of a photoactive electron donor-acceptor (EDA) complex between the substrates (Scheme 3). EDA is a molecular aggregate in the ground state, formed by the association of an electron-rich substrate with an electron-accepting molecule. Light excitation results in intramolecular single-electron-transfer (SET) event that occurs within the solvent cage, resulting in the formation of radical cation A and radical anion C. The radical cation A undergoes fast cyclopropane ring opening into β-ketoalkyl radical B which in turn reacts with C producing the cross-coupling products 3. In the reaction with benzylidene malononitrile 2a, cyclopropanol 1a produced 3aa in 25% yield after 3 h of UV light irradiation in acetonitrile. A significantly higher yield of 77% was achieved after 24 h of irradiation. Phenyl cyclopropanol 1g, possessing a less electron-donating aryl moiety, reacted with 2a noticeably slower, resulting in only a 13% yield of 3ga after 3 h. Similarly, the reaction of 1a with less electron-deficient dimethyl fumarate 2f also occurred with reduced rate, yielding 3af in only 16% yield after 19 h. These results stay in line with intermediacy of EDA complexes, which are stronger for more electron-deficient alkenes and for cyclopropanols with more electron-rich aryl moieties. |
64c3a8d8658ec5f7e5517a32 | 5 | To confirm the formation of an EDA complex between 1a and 2a, a UV-Vis spectroscopic study has been conducted (Figure , a). As commonly observed for weakly associated EDA complexes, a UV absorption band for a mixture of 2a and 1a (blue line in Figure , a) experienced a bathochromic shift with respect to the individual substrate 1a (gray line in Figure ). The excessive absorption in the range of ca. 370-400 nm can be more clearly visualized by subtracting the spectrum of 1a from the spectrum of the mixture (yellow line). To obtain additional evidence for the complex formation and to evaluate the association constant, a 1 H NMR titration experiment was performed (Figure ). Clear and gradual changes in the chemical shifts of the two diagnostic protons Ha and Hb of 2a were observed, shifting towards the highfield region of the 1 H NMR spectrum upon the addition of 1a (ranging from 1 to 95 equivalents). Fitting the experimental titration curve (see Figure in the Supporting Information) yielded the value of the association constant Ka = 0.16 M -1 . Scheme 3. EDA complex-mediated cross-coupling of aryl cyclopropanols with electrophilic alkenes. Reaction conditions: Alkene 2 (0.2 mmol), cyclopropanol 1 (1.5 equiv.), CH3CN (0.2 M), 370 nm Kessil LED, under Ar, r.t. Yields were determined by H NMR analysis of the crude reaction mixture against triphenylmethane as an internal standard. These results suggest that the EDA interaction is relatively weak, even between electron-rich cyclopropanol 1a and highly electron-deficient olefin 2a, though sufficient to engage these molecules into a photoinduced charge transfer process. Furthermore, the presence of an electron-rich aromatic moiety in 1a is probably responsible for the formation of the respective EDA complex, making it unlikely to be formed for aliphatic cyclopropanols. Consequently, electron transfer between the substrates is either accelerated or enabled by TBADT acting as a photoredox catalyst, and the catalytic process should likely generate the same key intermediates A and C. The reaction should start with a SET oxidation of cyclopropanol by photoexcited TBADT leading to the formation of radical cation A and its further rearrangement into β-keto alkyl radical B. The generation of B was supported by performing the reaction in the presence of TEMPO (2,2,6,6-tetramethyl piperidine-N-oxyl) as a radical scavenger. TEMPO interrupted the formation of 3aa as the TEMPO-trapped adduct 5 was formed instead (83% 1 H NMR yield, Scheme 4, a). In addition, the reaction with a disubstituted cyclopropanol 1s smoothly afforded 3sa (75% yield, Scheme 4, b), indicating that the ring opening occurred exclusively at the most substituted C1-C2 bond of the cyclopropane ring and therefore proceeded via a more stable secondary alkyl radical. After the addition of β-ketoalkyl radical B to the double bond of an alkene, HAT from solvent, cyclopropanol or other source might occur and result in the formation of product 3. Consequently, deuterium labelling experiments were performed to trace the origin of the δ-hydrogen in 3aa. First, D atom was not abstracted from acetonitrile-d3 (Scheme 4, c) which makes HAT from this solvent and from dichloromethane with a similar C-H bond dissociation energy of 96 kcal mol -1 unlikely. However, the reaction of benzylidene malononitrile 2a with deuterium-labeled cyclopropanol 1a-d (87% D) in dry dichloromethane resulted in 3aa-d with 69% D incorporation at δ-position, indicating that OH-group of 1a is the source of the δhydrogen (Scheme 4, d). More interestingly, the reaction of 1a-d was much slower than that of 1a, resulting in only 44% yield after 20 h of irradiation. The remarkable difference in the reaction rates along with lower deuterium incorporation in the product 3aa-d (69% D), in comparison to the content of deuterium in the starting material 1a-d (87% D), were clear consequences of a primary kinetic isotope effect (KIE) associated with the breaking of O-H bond in 1a. Moreover, the KIE was an indication that oxidation of 1a is the ratelimiting step of the whole process. However, it remained unclear whether a hetero-(SET followed by deprotonation) or a homolytic (HAT) cleavage of the O-H bond in 1a was involved. When cyclopropanol 1a underwent the reaction with dimethyl fumarate 2f in acetonitrile/D2O mixture (Scheme 4, e), rather fast (4 h) reaction occurred and produced 3af-d (98% D incorporation) in 87% yield. In this specific experiment, 3af served as a more reliable mechanistic probe than the previously used malononitrile adduct 3aa, as the latter compound experienced fast exchange of its more labile δ-H to deuterium with D2O (see section 4.3.3 in the Supporting Information). Based on these results, as well as considering high BDE of the O-H bond in alcohols and especially O-D bond in D2O (125 kcal mol -1 ), we concluded that the heterolytic cleavage of O-H bond and proton transfer are more likely than HAT. The conclusion was also bolstered by an observation that HAT processes occurred slowly under the standard reaction conditions, even in the case of relatively weak C-H bonds. For example, when ketone 6 was subjected to the reaction with benzylidene malononitrile 2a under the standard conditions (Scheme 4, f), HAT from the methoxy group occurred and the formation of product 7 was confirmed by H NMR and HRMS analysis, but in low 13% yield after 16 h. Furthermore, TBADT did not provoke any notable C-H activation side processes at sensitive benzylic sites during investigation of the substrate scope (for example, in the synthesis of compounds 3la and 3ma). A plausible mechanism, which stays in accord with the collected experimental data, is shown in Figure . Photoexcitation of TBADT followed by intersystem crossing produces the triplet excited state *[W10O32] 4-. The photoexcited W(VI) species oxidizes cyclopropanol 1 via SET, resulting in cation radical A. |
64c3a8d8658ec5f7e5517a32 | 6 | Disproportionation of the formed [W10O32] 5-regenerates the initial photocatalyst [W10O32] 4-and also produces W(IV) species [W10O32] 6-, which is a stronger reductant than initially produced [W10O32] 5-. The reduced forms of the photocatalyst turn the reaction mixture deep blue, which was also observed in our experiments. Considering the remarkable influence of an alkene as well as the result of D2O quenching experiment (Scheme 4, e), we believe that electron-deficient alkenes 2 act as one-electron oxidants to produce anion radical C, which recombines with β-ketoalkyl radical B to form anionic intermediate D. Upon protonation of D, product 3 is finally formed. We do not completely exclude an alternative mechanism which involves Giese-type radical addition of the nucleophilic radical B to electrophilic alkene, followed by SET from [W10O32] 5-or [W10O32] 6-to reduce the initially formed radical adduct (see Scheme S1 in the Supporting Information). -are reported by Ravelli et al. and realigned with respect to Fc + /Fc couple. To support the mechanism with additional experimental evidence, cyclic voltammetry (CV) measurements of the series of cyclopropanols and alkenes were performed. The values of peak potentials (Ep) have been compared, as all the CV experiments, including measurements for [W10O32] 5-/[W10O32] 6-and [W10O32] 4- /[W10O32] 5-pairs, were conducted under the same conditions (Figure ). Electroanalytical characteristics of TBADT have been extensively studied and provide information on the electrochemical potentials for single electron oxidation and reduction of the active catalytic species in acetonitrile and other solvents, but the values vary in the presence of water and are pH-dependent. Examination of the reaction scope (Scheme 2) showed that yields of products and the reaction rates drastically depend on nature of reacting cyclopropanols and alkenes. Such behavior could be correlated with the values of the respective oxidation and reduction potentials, as summarized in Figure . For example, 2-phenylethyl cyclopropanol 1l is the most resistant to oxidation (Ep = +1.3 V) among the three cyclopropanols tested, which agrees with its slow reaction rate (20 h) delivering 3la in a moderate 53% yield. This contrasts to fast and high yielding reaction of electron-rich 1a, which is the strongest reductant (Ep = +0.84 V). The reactions of the least susceptible for SET oxidation aliphatic cyclopropanols are accelerated or even initiated by the addition of an external oxidant (NFSI or K2S2O8). The role of the oxidant remains unclear, however substoichiometric loadings (20-50 mol%) imply that it could act as a radical initiator that induces homolytic (rather than via SET oxidation) cleavage of the O-H bond in 1 or could recover the reduced forms of TBADT catalyst formed in parasitic C-H activation processes. An analogous correlation between reactivity and single-electron oxidative properties has also been observed for alkenes: stronger oxidants, as revealed by comparison of the respective Ep values (Figure , c), are more favorable substrates in their reaction with 1a. The clear dependence on the electronic properties of the alkenes offers further support for the intermediacy of the anion radical C. It was noted that the reduction potentials measured for the radicals formed in the Giese reaction showed these are much stronger one-electron oxidants (-0.3 to -0.1 V vs Fc + /Fc) and the reaction rate in this case should be less dependent on starting alkene. Moreover, we believe that disproportionation of [W10O32] 5-delivering a stronger W(IV) reductant [W10O32] 6-is essential to enable the reduction of 2 into C. |
64c3a8d8658ec5f7e5517a32 | 7 | In conclusion, we have developed a general decatungstate-photocatalyzed methodology for the redoxneutral alkylation of electron-deficient alkenes by β-ketoalkyl radicals derived from readily available tertiary cyclopropanols. A broad range of aryl-substituted cyclopropanols and electron-poor alkenes are favorable substrates in this protocol, which has been adjusted to involve less reactive alkyl cyclopropanols by the addition of NFSI as a substoichiometric additive. The developed protocols were conveniently scaled up by translating them into a continuous flow regime, allowing the preparation of synthetically valuable distantly functionalized ketones in gram quantities. Mechanistic studies confirm the role of TBADT as a photoredox catalyst that facilitates single electron transfer between the reactants. The formation of EDA complexes between electron-donating aryl cyclopropanols and electron-accepting alkenes has been disclosed. The EDA complexes can transform into the same distantly functionalized ketone products by means of photo-induced charge transfer, although the reaction is slower than the TBADT-catalyzed process. |
64e48388dd1a73847f4a29d3 | 0 | The component of cell death produced by TS inhibition is abnormal uracil incorporation into DNA resulting from build-up of the cytotoxic nucleotide intermediate dUTP and the dUTP pools accumulate. DNA polymerases use dUTP in place of TTP during DNA synthesis and activating repeated rounds of uracil repair due to continue dTTP depletion, resulting in extensive DNA damage and cell death. The enzyme deoxyuridine triphosphate nucleotidohydrolase (dUTPase), encoded by the DUT gene, which hydrolysing dUTP to dUMP, thereby eliminating dUTP from the DNA biosynthetic pathway. Even though the depletion of dTMP and the accumulation of dUMP is common in cancer cell lines undergoing TS inhibition ,the ability to accumulate dUTP varies definitely. The dUTPase could protect colorectal and breast cancer cells from cytotoxicity induced from dUTP pool expansion and subsequent uracil misincorporation into DNA induced during TS inhibition. Also dUTPase over expression in tumour specimens is associated with resistance to 5-FU in colorectal cancer but numerous solid tumours over express dUTPase, cancel the potential for dUTP pool expansion and cytotoxicity induced by the uracil-DNA misincorporation pathway. After all, if dUTPase activity could be blocked in the presence of TS inhibition and the combined cytotoxicity from concomitant TTP pool depletion, along with uracil misincorporation would result in enhanced DNA damage and cell death. |
64e48388dd1a73847f4a29d3 | 1 | The , PPP produces nicotinamide adenine dinucleotide phosphate (NADPH), which serves as a hydrogen donor in various biosynthetic processes and has an important role in case of oxidative attack by the infected host. It has also been recognized as an attractive drug target ,and one of its key enzyme is glucose-6-phosphate dehydrogenase (G6PDH). After all, only a limited number of studies are available on enzymes of the PPP (e.g., for parasites and for mammalian homologs, although this pathway is of particular importance. Besides, three-dimensional (3D) structures of the second and third PPP enzymes of T. brucei, the 6-phospho-gluconolactonase (6PGL), and the 6-phosphogluconate dehydrogenase (6PGDH) have been solved. This is particular interest because structural differences between the parasitic and mammalian 6PGDH have been exploited to design molecules that would be specific inhibitors of the T. brucei enzyme. |
64e48388dd1a73847f4a29d3 | 2 | For developing any natural product for clinical application we need to comprehensively understand and identify its molecular targets and mode of action. The rapid advancement of high throughput screening, structural elucidation and combinatorial synthesis have revitalized the potential of plant derived compounds as chemotherapeutic agents against cancer, and Computational screening programs such as molecular docking has greatly helped in rapid screening of chemical entities against their macromolecular targets(Gurung, Bhattacharjee and Ali, 2016). Such as in silico toxicity screens have routinely been employed in drug discovery pipelines to study the pharmacokinetic and pharmacodynamics properties of the selected drug like compounds before proceeding to experimental trials. |
64e48388dd1a73847f4a29d3 | 3 | Molecular interactions including protein-protein, enzyme-substrate, protein-nucleic acid, drugprotein, play important roles in many essential biological processes, such as signal transduction, cell regulation, gene expression control, enzyme inhibition, antibody-antigen recognition. And these interactions very often lead to the formation of stable protein-protein or protein-ligand complexes that are necessary to perform their biological functions. . The tertiary structure of proteins is necessary to understand the binding mode and affinity between interacting molecules, and it is often difficult and expensive to obtain complex structures by experimental methods, such as X-ray crystallography or NMR. So, docking computation is considered an important approach for understanding the protein-protein or protein-ligand interactions , and the protein structures determined by structure databases such as Protein Data. Bank (PDB) and Worldwide Protein Data Bank (wwPDB) have over 88000 protein structures. Molecular docking is a widely used computer simulation procedure to predict the conformation of a receptor-ligand complex, and the accurate prediction of the binding modes between the ligand and protein is of fundamental importance in modern structure-based drug design. where the receptor is usually a protein or a nucleic acid molecule and the ligand is either a small molecule or another protein, the most important application of docking software is the virtual screening, in which the most interesting and promising molecules are selected from an existing database for further research. |
64e48388dd1a73847f4a29d3 | 4 | In silico computational protein-Ligand docking has become an important tool for drug discovery and development and is generally initiated with structure identification of known target protein molecules of medical interest. Molecular docking is then used to predict protein interactions with candidate small molecules (ligands) according to conformations and binding free energies, Which are expressed as ligands-protein binding forces in kilocalories per mole (kcal/mole) .As a direct and rational approach for drug discovery , and computational docking analysis establish virtual models of ligand-protein interactions at the atomic level and can be used to inform subsequent validation using traditional in vitro and in viva assays, thus minimizing the time and cost of the drug discovery process .VS, an in-silico HTS method, consists of virtually placing (docking) collections of millions of compounds into a biological target . Thus in basic requirements for VS are: |
64e48388dd1a73847f4a29d3 | 5 | In this study, we analysed the protein ligand interaction highest binding affinity and compare their properties with Lipinski rule of five. For this study, 500 ligands were collected from Dictionary of marine natural product and their properties compare with Lipinski rule of five. From 500 ligands, 356 ligands follow the Lipinski rule of. And for this protein ligand complex two protein were collected from protein data bank (PDB). |
64e48388dd1a73847f4a29d3 | 6 | The drug discovery process is a very complex and includes an interdisciplinary effort for designing effective and commercially feasible drug, Combination of rational drug design and structure biology leads to discovery of novel therapeutic agents. In pharmaceutical, medicinal as well as in other scientific research ; a computer plays a very essential role, even in development of new compound in quest for better therapeutic agents and for this purpose computer aided drug design (CADD) centre work with collaboration between structure biologists, biophysicists and computational scientists for discovery of new chemical entities . |
64e48388dd1a73847f4a29d3 | 7 | Molecular docking helps us in predicting the intermolecular framework consists of between a protein and a small molecule and protein and suggest the binding condition responsible for prohibition of the protein. Two molecules can bind in a number of ways the binding of a protein and protein or a protein and small molecule . |
64e48388dd1a73847f4a29d3 | 8 | Molecular docking is a widely used computer simulation procedure to predict the conformation of a receptor-ligand complex, where the receptor is usually a protein or a nucleic acid molecule and the ligand is either a small molecule or another protein . Where the protein are selected from Protein Data Bank (PDB). In this work, the molecular docking calculations were performed using the AutoDock program and AutoDock program has become widely used due to its good precision and high versatility; moreover, the latest version of AutoDock (version 4.0) added flexible functions to the side chains in the receptor. |
64e48388dd1a73847f4a29d3 | 9 | Drug discovery is the procedure when new applicant drug are discovered and classically drugs were discovered by identifying the active substance from classical therapy as with penicillin. Discovering drugs that may be a commercial success, or a public health success, includes a complex interaction between investors, industry, academia, patent laws, regulatory exclusivity, marketing and the need to balance secrecy with communication. |
64e48388dd1a73847f4a29d3 | 10 | Drug design frequently but not necessarily depends on computer modelling techniques ,this type of modelling is sometimes referred to as computer-aided drug design and in most of the case, drug design includes the design of molecules that are complementary in shape and charge to the bio molecular target with which they interact and therefore will bind to it. Although design techniques for prediction of binding affinity are reasonably successful, there are several other properties, such as bioavailability, metabolic halflife, side effects, etc., that first must be optimized before a ligand can become a safe and efficacious drug In addition to small molecules, biopharmaceuticals including peptides and especially therapeutic antibodies are an increasingly important class of drugs and computational methods for improving the affinity, selectivity, and stability of these protein-based therapeutics have also been developed. |
64e48388dd1a73847f4a29d3 | 11 | Most in silico DTBA prediction procedure developed to date use 3D structural information, DTBA(drug target binding affinity) signs the strength of the interaction or binding between a drug and its target which was demonstrated to successfully contribute to the drug design. The 3D structure information of proteins is used in the molecular docking analysis which followed by applying search algorithms or scoring functions to comfort with the binding affinity predictions and this whole process is used in the structured-based virtual screening. The benefits of formulating drug-target prediction as a binding affinity regression task, is that it can be transformed from regression to either binary classification by setting specific thresholds or to ranking problem Even though the macromolecular structure and small molecule databases are an essential factor for drug discovery in silico, but it still needs the effective software for performing virtual screening on targets and small molecule databases, the computer-aided software has successfully applied into virtual screening studies and accelerated the process of drug discovery. Usually, the computational methods for drug discovery can be divided into ligandbased (indirect) and structure-based (direct) technique is the ligand-based drug design methods contain quantitative structure-activity relationship (QSAR), pharmacophore, etc. and the structure-based drug design contains molecular docking. The computer-aided software has successfully applied into virtual screening studies and accelerated the process of drug discovery. AutoDock, has faster run speed and more accurate binding mode predictions than AutoDock62. MD software can be used to study the dynamical interaction between targets and ligands at the atomic level. It can profile more detail and accurate interaction information for ligands in the pocket of receptor than molecular docking , and this MD simulations have become the popular method to study the mechanism of activated, inactivated states and ligand interaction on different targets, such as ion channel receptors, etc. |
64e48388dd1a73847f4a29d3 | 12 | One choice of drug design using high throughput examinations and computational tools is that it can largely lower the use of animals in activity testing, moreover, in vitro experiments complemented with computational methods have been largely used in early drug discovery to select compounds with more favourable ADME and toxicological profiles naturally, drugs are organic small molecules produced through chemical synthesis, but biopolymer-based drugs (also known as biopharmaceuticals) over through biological processes are becoming increasingly more common. |
64e48388dd1a73847f4a29d3 | 13 | Based on the many protein targets of diseases, multiple approaches including biological or in silico had been designed to screen new drugs towards the diseases treatment and these approaches are, very expensive and demand a great deal of background knowledge. where in pharmacology, many chemical libraries of synthetic small molecules, natural products or extracts were screened in vitro or in vivo , such as intact cells, whole organisms or cell-free systems to identify substances that have a desirable therapeutic effect and the recent experimental process to approach drug discovery includes several well defined biochemical examinations to screen those compounds that can interact or bind with certain binding partners, such as receptor/ ligand binding analysis(Hulme and Trevethick, 2010)enzyme-activity evaluation. For protein binding studies different technique can be used such as X-ray crystallography structural analysis NMR, , calorimetric, affinity chromatography, and protein mass spectrometry which are common strategic tools in drug discovery and these techniques aim to detect the separation of compounds from the studied proteins and monitor changes in intrinsic parameters of the targets upon forming a complex with tested drugs. |
64e48388dd1a73847f4a29d3 | 14 | Compared to in vitro screening, in vivo testing is better suited for observing the overall effects of an examination on a biological subject even though vitro assays can sometimes yield misleading results with drug candidate molecules that are irrevalent in vivo, efficacy verification in vivo is especially crucial in drug discovery process. |
64e48388dd1a73847f4a29d3 | 15 | In spite of various models developed for drug discovery, but most therapeutic drugs still fail in clinical tests and one of the reasons is attributed to sufficient clinical predictive power of our current model systems. usually, for a drug discovery research in the laboratory, in combination with the chemical examinations, cell-based and in vivo testing would perform more efficiently to obtain effective lead compounds for other drug development, although the high genetic similarities between human and mice, physiological differences affect the course of disorder in mice models when some genetic disorders in human do not have the same symptoms in mice. |
64e48388dd1a73847f4a29d3 | 16 | Drug-likeness widely used to describe physicochemical properties of ligands that are orally bioavailable and therefore the compounds should be biologically active ). An examination of drug-likeness in the condition of oral bioavailability was first reported by Lipinski, Lombardo, Doming, and Feeney in a pivotal analysis which determined the physicochemical properties common to the 90 Th percentile of almost 2500 drugs and candidate drugs reaching Phase II clinical trials .Lipinski defines 'drug-like' compounds that have good absorption, distribution, metabolism and excretion profiles to pass through the clinical trials. He summarized the physicochemical properties present in a drug-like molecule as rule of five. The rule is essential to keep in mind during drug analysis at the same time pharmacologically effective lead structure in develops and step-wise to increase the activity and also selectivity of the compound as well as to confirm drug-like physicochemical properties are managed as described by Lipinski's rule. Applicant drug that ensured to the RO5 move to have lower attrition rates when clinical trials and hence have an increased chance of reaching market. |
64e48388dd1a73847f4a29d3 | 17 | Lipinski's rule of five also known as the Pfizer's rule of five or simple rule of five (RO5) is rule of thumb to assay drug-likeness or regulate if a chemical compound with a positive pharmacological or biological activity has chemical properties and physical properties and would make it a likely orally active drug in humans. The rule was formulated by Christopher A. Lipinski in 1997, based on the observation that most orally administration drugs are relatively small and slightly lipophilic molecule. The rule states that poor absorption or infiltration of drug more reasonable when the chemical structure fulfils two or more of the following criteria: |
64e48388dd1a73847f4a29d3 | 18 | PyMol and Discovery studio biovia 2019 were employed to visualize and modify receptor and ligand structures and Autodock vina, PyRx was the primary docking program used in this study. The protein-ligand structure was determined by PyMol, and the ligand-protein complex structure and their different binding site were visualized by biovia studio. |
64e48388dd1a73847f4a29d3 | 19 | There 500 ligands were collected from Dictionary of Marine Natural products, which compare with Lipinski rule of five. Many natural products occupy chemical space which is different from and complementary to those in synthetic libraries, making natural product libraries attractive for drug discovery. Actually, natural products and their caused have been a main source of pharmaceutical leads and therapeutic agents, specifically , marine natural products (MNPs) have indicated exceptional efficacy and potential as anticancer therapeutics . |
64e48388dd1a73847f4a29d3 | 20 | The type of ligands chosen for docking will depend on the goal and it can be obtained from various databases, such as ZINC or/and PubChem, or it can be sketched by means of Chemsketch tool . For structure preparation ChemAxon MarvinSketch was used and different physicochemical properties of the ligands were collected from ChemSpider database All ligand compound saved as ligand01, ligand02, ligand03, ligand04 etc. |
64e48388dd1a73847f4a29d3 | 21 | • dUTpase playing an essential roles in cellular nucleotide metabolic process, it catalyze the electrolyze of dUTP to dUMP and pyrophosphate (PPi). • Firstly, the enzyme supplies dUMP to synthesize dTTP ,a DNA building block, and secondly, the enzyme lowers the dUTP/dTTP ratio in the cell, so preventing the misincorporation into genome DNA . • And deficiency of the enzyme would result into too high uracil content in DNA,The thymine-less cell death due to overdosed DNA repair activity • Because of their essential role, it has been suggested that as potential drug targets for the therapy of cancer, and infectious diseases. |
64e48388dd1a73847f4a29d3 | 22 | • Several chemotherapy drugs such as 5-fluouracil treat neoplastic disease, including head and neck cancer, breast cancer and gastrointestinal cancer by targeting TS in thymidylate metabolism. • High levels of DUT-N have been combined with chemo resistance and faster tumour progression, and thus, could also serve as a diagnosis marker for overall survival and response to chemotherapy. • Correspondingly , DUT is naturally overexpressed in hepatocellular carcinoma and may serve as a prognostic marker for the cancer . |
64e48388dd1a73847f4a29d3 | 23 | • 6-phosphogluconolactonase produces the alteration of 6-phosphogluconolactone to 6phosphogluconic acid where both medium in the oxidative phase of the pentose phosphate pathway and the glucose is converted into ribulose 5-phosphate. • The oxidative phase of the pentose phosphate pathway releases CO2 and results in the generation of two equivalents of NADPH from NADP+ and the final product, ribulose 5-phosphate, is more processed by the organism during the non-oxidative phase of the pentose phosphate pathway to synthesize biomolecules including nucleotides, ATP, and Coenzyme A. |
64e48388dd1a73847f4a29d3 | 24 | • Trypanosoma brucei is the causative agent of African sleeping sickness which is a lethal disease caused by the protozoan parasite Trypanosoma brucei . • These parasites have been the object of many studies in the current decades, and great progress has been made in understanding their biochemistry and these efforts have led to the development of treatments that are of limited efficiency and of great toxicity. |
64e48388dd1a73847f4a29d3 | 25 | Molecular docking was performed to using the Autodock vina program. During docking first AutoDock Vina, PyRx the target macromolecule is selected and it made in automatically pdbqt format when the protein made macromolecule. Then, the ligand is imported one by one, and then selected pdbqt protein and then click on start, the ligand automatically showed and selected it, also selected protein molecule. Then the docking is running and the result is produced. The negative values for binding affinity (or binding free energy) indicate that the ligand is predicted to bind to a target macromolecule; the more negative the numerical values for the binding affinity, the better is the predicted binding between a ligand and a macromolecule. |
64e48388dd1a73847f4a29d3 | 26 | DUTpase protein, the correlation between binding affinity and molecular weight of the ligands are given below in table From figure the different bonds are showed such as the pink colour is pi-donor bond, green colour is conventional bond and light green colour is carbon bond. There were also determined, pi-bond distance is 4.74, conventional bond is 3.00 and carbon bond is 3.23. |
64e48388dd1a73847f4a29d3 | 27 | Virtual screening using molecular docking programs has become a more popular compare to the development of new drugs and monetary costs of in silico drug screening analysed with traditional laboratory experiments . Molecular docking has various strengths ,that the method's ability to screen large compound databases at lowpriced compared to experimental techniques such as HTS is particularly notable . |
64e48388dd1a73847f4a29d3 | 28 | We have worked on correlation between physicochemical properties of protein-ligands and binding affinity to protein targets by virtual screening. The virtual screening is an important tool, which widely used as a computational high throughput screening. In this research we showed higher binding affinity and also compare with drug like properties. According to 'rule of five' which molecule contain molecular weight is less than 500, logP is less than 5, H-bond acceptors less than 10 and H-bond donors is less than 5, these molecule show higher binding affinity and show drug-like properties. |
64e48388dd1a73847f4a29d3 | 29 | In this study, we docked 500 ligands with 2 proteins, to identify highest binding affinity and also visualized their different chemical bonds with amino acids by Biovia study. All ligands, we compared with Lipinski rule of five, where 356 ligands follow Lipinski rule of five and docked with these two proteins by Autodock Vina, PyRx. From all these ligands, ligand 243 showed highest binding affinity between these protein-ligand complexes. The dUTpase protein and ligand 243 complex binding affinity is 18.2 Kcal/mole, where 6-PGL protein with ligand 243 complex binding affinity is -17.5 Kcal/mole. These protein-ligand complexes also showed different amino acid and their chemical bond such as conventional hydrogen bond, Pi-donor bond, carbon bond, water hydrogen bond etc. |
64e48388dd1a73847f4a29d3 | 30 | The dUTpase protein can be used as future development of drug like as anticancer therapy and for more therapeutic uses. And another protein 6-PGL can also be used as sleeping sickness and antimalarial therapy. It can be concluded that these derivatives could be used as a template for the future development through modification to design more potent therapeutic agents. |
61e53beceab6ef20aae41719 | 0 | Retrosynthetic analysis, first developed decades ago, plans the synthetic routes of pharmaceuticals and industrial chemicals by transforming a target molecule into simpler precursors until available building block molecules are reached. Computer assisted synthesis planning (CASP) has gained great interests since it accelerates the development of retrosynthetic routes. The emergence of chemical reaction databases and the ability to mine data from historical literature and patents made it possible for the development of numerous data-driven CASP methods. Recent successes in this field include: (i) generation of reaction networks based on graph theory, using vertices to represent molecules and directed edges to represent reactions from reactants to products, to understand connectivity among molecules, (ii) manual curation and algorithmic extraction of reaction rules and templates from historical reactions to predict functional group transformations, (iii) template-free deep learning methods to learn from historical reactions and plan new transformations, etc. |
61e53beceab6ef20aae41719 | 1 | A seemingly parallel development is that of synthetic biology, in which cells are engineered to produce target molecules. Synthetic biology is increasingly drawing attention of organic chemists for two reasons: (i) it can improve routes redox efficiency by finding metabolic shortcuts for the key synthetic steps, and (ii) most enzymatic reactions are highly selective. In addition, biochemical reactions are performed under mild operating conditions and usually with benign solvents, which may lower operational costs and reduce life cycle impact of syntheses. Similar to the reaction network of organic synthesis, a map to visualise synthetic biology space has been summarised to guide the biosynthetic planning. A number of pharmaceutical ingredients and bulk chemicals have been produced economically through metabolic engineering approaches. With the knowledge of molecular transformations from organic chemistry and synthetic biology, retrosynthesis relies on multi-step decision making to select optimal reaction routes among all feasible molecular transformations based on criteria such as exergetic efficiency, Efactor, and etc. In linear reaction routes, which include only one-to-one (reactant(s)-product only) 'wiring' (using the networks' jargon) of the reactions, the decision making could be done through exhaustive search of all possible reaction routes, and ranking of the routes based on a set of pre-defined criteria. However, in a topological-tree-styled reaction routes, which include multiple-to-multiple wiring reactions (including co-reactants and bi-products), the number of routes options increase exponentially with the increase in the number of synthetic branches and depth; exhaustive search becomes computationally expensive. In order to improve the efficiency of route design the machine learning method of reinforcement learning has been proposed for the application in synthesis planning. Reinforcement learning (RL) mimics how an intelligent 'decision maker' takes multi-step actions within a specific problem environment to maximise/minimise the cumulative rewards/penalties of the actions. In synthetic planning, the selection of each reaction step within a path is a decision making step. With the given rules and criteria costs, a 'decision maker' starts synthetic planning by trail-and-error, and algorithmically learns from the simulated experience to perform better in next iteration (an 'episode' within the RL jargon). |
61e53beceab6ef20aae41719 | 2 | For example, Schreck et al. trained a policy model to understand the potential costs of candidate reactions computed from reaction templates at a certain synthetic depth, and to select retrosynthetic pathways based on costs. The method was compared with a decomposition heuristic method 2 to prove its ability for synthetic planning. Similarly, in metabolic engineering, Koch et al. presented a code named RetroPath RL, which uses Monte Carlo Tree Search (MCTS) reinforcement learning to rank metabolic reaction rules to enable the development of biosynthetic routes. |
61e53beceab6ef20aae41719 | 3 | Despite the interest in chemoinformatics to combine multiple datasets to have comprehensive understanding of the chemical space, to the best of our knowledge, none has worked on combining organic chemistry and synthetic biology databases and understand the benefits from both in retrosynthesis planning. In this work, we mined a section of historical reactions from Reaxys ® database, 31 and all metabolic reactions from Kyoto Encyclopedia of Genes and Genomes (KEGG), which is an open-source manually curated bioinformatic library. We compared the influence of the presence of organic synthesis and synthetic biology past reaction data in a dataset used for identification of retrosynthesis pathways of a curated set of drug molecules, which were believed to be difficult to synthesise. To evaluate the identified routes we used atom economy, the number of reaction steps and price of molecular building blocks as key quantifiable performance criteria. Reinforcement learning method from Schreck et al. was adapted to build policy models to guide the search for retrosynthesis pathways. Different from other CASP tools, the synthetic pathways from this method were not assembled from predicted reactions (i.e., using reaction template or the algorithm generated reactions ), but using historical published data. This reduced the operational uncertainty over the identified paths to enable us to focus on the key research question of this work -how much added value could synthetic biology bring to synthetic organic chemistry in multi-step syntheses? |
61e53beceab6ef20aae41719 | 4 | The molecular building blocks of a target molecule are smallest precursors required to build up the structure of the target molecule. To make a collection of building blocks and their prices for synthetic planning, here we defined building blocks to be commercially or naturally available small molecules. The collection of commercially available building blocks came from 'buyable' molecules crawled from ChemSpace, 33 which is an online catalogue of small molecules. For most buyable molecules listed in ChemSpace, more than one price is given from multiple suppliers. We only chose the lowest price available for these molecules. |
61e53beceab6ef20aae41719 | 5 | The naturally available building blocks are freely available cofactor metabolites from cell organism in metabolic reactions, such as ATP and NADPH, and a list of such molecules curated by Blaβ et al. was used for the naturally available building blocks. The price for these naturally available molecules is zero. In the enzyme-based industrial processes, although the natually occuring molecules are free to acquire, some of these molecules, specifically co-factors, are difficult to recover and recycle, which makes them economically unviable. The common industrial solutions include stoichiometric design that balances each cofactor occurring in the total pathway, or integrating multiple pathways to link the generation/degradation of cofactors. For simplicity, in this work we only consider the acquisition price of the building blocks to demonstrate the overall approach; the cost of separation and other process 'costs' of the syntheses are deliberately left out from the current study. |
61e53beceab6ef20aae41719 | 6 | Figure shows the price distribution of all building block molecules. Price of approximately 1/10 th of all molecules from ChemSpace range from 10 ! to 10 " USD/g, which is unreasonably expensive. Therefore, these molecules were removed from the set of building blocks to lower the costs of the potential synthetic routes. In total, we selected 24,282 commercially and 451 naturally available building blocks respectively. A full list of building block molecules can be found in the Supporting Information (SI). |
61e53beceab6ef20aae41719 | 7 | A section of reactions was mined from Reaxys ® . This is the same section as the molecules and reactions to generate the network of organic chemistry (NOC) from our previous studies. All reactions were mined from KEGG reaction database. Since in most cases enzymes catalyse metabolic reactions from both directions, all reactions were assumed to be reversible. |
61e53beceab6ef20aae41719 | 8 | From the two datasets, all molecules and reactions were recorded with their own identification numbers. Due to the use of different identifiers, to merge both datasets we used RDKit package to pairwise compare molecular canonical SMILES strings for all molecules, and reaction SMARTS for all reactions in both datasets. All KEGG molecules and reactions found in Reaxys were renamed with Reaxys identifier in the local datasets. An example of these molecules and reactions can be found in SI. The statistics of both datasets are shown in Figure . Data from KEGG is significantly more sparse compared with the dataset of reactions mined from Reaxys. Among KEGG data, a proportion of molecules and reactions overlap with the Reaxys data, since Reaxys includes mined reaction data regardless whether they are from organic synthetic or bio-synthetic sources. To compare the optimal reaction routes computed from candidate reactions from different sources, three local reactions pools were created: reactions from KEGG were labelled as a biological reactions pool (green + brown in Figure ), those from Reaxys and excluding the intersection between Reaxys and KEGG, were labeled as chemical reactions pool (pink), whilst the union of the two sets became the hybrid reactions pool (pink, green + brown). The visualisation of the hybrid reaction network in node and edge representation is shown in Figure . |
61e53beceab6ef20aae41719 | 9 | Since KEGG reactions were manually curated, by no means could the reactions cover the entire synthetic biological space. In reality, the intersection between Reaxys and KEGG may be larger than the brown area. Thus, the chemical reactions pool (pink) possibly still includes metabolic reactions. However, by far boundary between the chemical and biological datasets has been set from the best of knowledge. |
61e53beceab6ef20aae41719 | 10 | We noticed that in Reaxys multiple molecules may sometimes share one identical canonical SMILES string, which gives database noise or high order structure differences (isotopes, stereoisomers, etc.); in the intersection of the two datasets one KEGG molecule may have multiple counterparts in Reaxys. The statistics in the intersection of Venn diagram were counted based on the data from the KEGG-extracted dataset to avoid this issue. |
61e53beceab6ef20aae41719 | 11 | In reinforcement learning multi-step decision making, each decision making step is associated with a reward/penalty, whilst the whole multi-step decision making process is associated with an expected return/cost, which is the acummulation of rewards/penalties of all decision-making steps. For each of the decision making steps, having multiple options, a well-trained policy model predicts expected returns/costs of the options by foreseeing the cumulative return/cost of the following steps from the current options, and then selects the option with maximised expected return, or the minimised expected cost. In the synthetic problem, each decision making step is to select a reaction option from the reactions pool for the current molecule, and the whole multi-step decision making process is to assemble the multi-step reaction pathway. Here to assess the reaction options, we chose the penalty-expected cost evaluation system and designed reaction assessment scores to represent the penalty to select the candidate reactions. The scores would later be also used to quantify performance of synthesis planning from the three different reactions pools. |
61e53beceab6ef20aae41719 | 12 | Various criteria could be applied to evaluate candidate reactions, subject to the optimisation objectives and data availability. In this work, our objective was to find efficient reaction routes whilst maintaining environmental efficiency. Although many reactions in Reaxys provide attributes such as yield, selectivity and reaction conditions, such information cannot be found from the KEGG dataset or other commonly used biological databases. Therefore, only global criteria determined from the data available in both Reaxys and KEGG databases were used to design the assessment scores. After trials to avoid failure of computation and biases, the global criteria were designed to include atom economy of reaction steps and price of building blocks to consider both route efficiency and operational costs. Of course, changes in global criteria would significantly alter the optimisation results. Here we chose a minimum set of criteria to demonstrate the overall approach. |
61e53beceab6ef20aae41719 | 13 | The design of penalty based on the global criteria was adapted from Schreck et al. For a reaction pathway, penalties were added to the reactions and the building block molecules. For any reaction or building block molecule in the pathway, the penalty was designed to be lower than 1. The penalty of a reaction is shown in Eq. 1. |
61e53beceab6ef20aae41719 | 14 | The penalty is lower when the atom economy of the desired product in the reaction is greater, where the atom economy (𝐴𝐸 # ) counts for the ratio of desired products 𝑖 over all products in a reaction step. Due to unavailability of reaction stoichiometry in Reaxys, atom economy was determined on the basis of molecular weights (Eq. 2). |
61e53beceab6ef20aae41719 | 15 | Apart from building blocks, the reaction pathway may also terminate at a dead-end molecule, which means no other reaction link with the molecule, or a maximum-depth molecule, which means the end point molecule reaches maximum allowed route depth from the target molecule, which was set to be 10 synthesis steps. The 'decision maker' fails to find a proper pathway in these cases, and therefore, adapted from Schreck et al., the penalty for a dead-end molecule is 100, and the penalty for a maximum-depth molecule is 10. |
61e53beceab6ef20aae41719 | 16 | The expected cost of a molecule in the reaction pathway is the cumulative penalties of all reactions and end-point molecules from the sub-pathway from the molecule as target molecule to its sub-branches, (Eq. 3). The expected cost of a molecule is also equal to the penalty of the reaction linked with the molecule as a product, plus expected costs of all reactants in the reaction. For example, in Figure , the expected cost of 𝑚 / is the sum of penalty of a building block 𝑚 " , and a max-length molecule 𝑚 01 , plus the penalty of reactions 𝑟 0 and 𝑟 ! . |
61e53beceab6ef20aae41719 | 17 | A reinforcement learning approach adapted from Schreck et al. was applied to create policy 'decision makers' towards identification of the optimal reaction routes. Optimisation workflow shown in Figure was conducted in 20 iterations to optimise the decision-making process and provide promising policy models. The policy model was later used to suggest near-optimal reaction routes for target molecules from the reactions pool. To compare the impacts of |
61e53beceab6ef20aae41719 | 18 | Eq. 3 chemical, biological and hybrid reactions pools, the same workflow was run three times to train and produce three policy models from the three reactions pools respectively. includes: (i) a reactions pool compromising all molecules and reactions for the 'decision maker' to choose from, (ii) evaluation score functions to assess reactions and synthetic routes, and (iii) a set of molecules as target molecules to initialise retrosynthesis planning. |
61e53beceab6ef20aae41719 | 19 | For the target molecule set, molecules were filtered to be in the SMILES string length of 20 to 100. This was to maintain the target molecules from different datasets with fair synthetic difficulty. The aim was to include only 100,000 molecules to maintain reasonable computational costs. This was the case for the chemical and hybrid reactions pools. 100,000 molecules (excluding the molecular building blocks) were randomly selected from the molecule set as targets. Also, in each iteration of the optimisation, the target molecule set was reshuffled to increase randomness. However, since the biological dataset records only approximately 30,000 molecules, all molecules with SMILES string length of 20 to 100 (building blocks exclusive), i.e. 12,281 molecules were included as the set of target molecules to compute synthesis planning. |
61e53beceab6ef20aae41719 | 20 | For each target molecule, to compute its retrosynthesis route, all reactions in the reactions pool using the target molecule as one of the reaction products were marked as possible reaction options. If no reaction was found from the dataset, the molecule was marked as a dead-end molecule, as no synthesis step could be further added to the molecule. A dead-end molecule in the pathway is highly disfavoured by the 'decision maker'. |
61e53beceab6ef20aae41719 | 21 | A 'decision maker' selected one of the reaction options as the next synthesis step for the target molecule. For each reactant in selected reaction, as shown in Figure , if the reactant was a building block or a dead-end molecule, no further synthesis step is required. If not, the reactant became the next step target molecule. The same procedure was repeated to add the next reaction to the retrosynthesis route until all end-point molecules at all branches (resulted from multiple reactants reactions in the route) were building blocks, dead-end molecules, or maximum-depth molecules, where the maximum allowed depth was set to be 10 synthesis steps from the target molecule, which is also highly disfavoured. shown in red on top of the nodes. In the schematics, all reactions have penalties of 1 and all building blocks have no penalty only for simplification. However, in most cases, reactions and building blocks always have penalties ranging from 0 to 1. |
61e53beceab6ef20aae41719 | 22 | Random sampling brought noises to the decision-making problem, which randomly selected over all reaction options, and made it possible to explore over the decision space. Using the trained policy model, the expected costs of all reactants based on their molecular fingerprints and residual depths (discussed below) were predicted for each reaction option. The policy model 'decision maker' selects the candidate reaction 𝑟 ) , which has the minimum sum of predicted expected costs of all reactant molecules (Eq. 5). Essentially, this means the policy model would understand the expected costs of molecular structures after exploring the reaction space, and always point to the molecule structures that are easier to synthesise as the following steps. |
61e53beceab6ef20aae41719 | 23 | Thus, the 'decision maker' started with random sampling to learn from trial-and-error. From iteration 1, the possibility to use the updated policy model 'decision maker' gradually increased, and from iteration 10, the reaction pathway was optimised only from the policy model. Suggested by Schreck et al., not only does the expected cost of a molecule depend on the molecule itself, but also the depth of the molecule in the pathway. If the molecule requires a long synthesis pathway, whether the pathway of a molecule reaches maximum-depth molecules or building-block molecules relies on its residual depth, i.e. the maximum allowed depth (10 steps) subtracted by the current depth from the target molecule. To learn from the simulated experience, the residual depth and the corresponding expected costs of all molecules in the pathway were collected following the designed penalty rules. This did not include the side- |
61e53beceab6ef20aae41719 | 24 | Same procedure was repeated for all target molecules to collect residual depths and expected costs of all simulated molecules. For the former 10 iterations, the simulation of each target molecule was repeated 10 times to add randomness to the built pathways. However, for the latter 10 iterations, since all pathways were built by the trained policy models, the repeated simulation results were identical, and thus, only one simulation was required for each target molecule. In each iteration, approximately a million expected cost values of molecules at their corresponding residual depths were collected. The multiple expected costs of same molecule at same residual depth were averaged to count into the training data. |
61e53beceab6ef20aae41719 | 25 | As shown in Figure , at each iteration, the trained MLP (discussed below) was eventually updated as the policy model 'decision maker' for the next iteration. The optimisation was terminated after 20 iterations, and the policy model at the last iteration became the final 'decision maker' to predict expected costs of molecules, and select reactions based on Eq. 5 to build retrosynthesis pathways. |
61e53beceab6ef20aae41719 | 26 | Following Schreck et al., machine learning models mimic the mathematical relationship between the molecules and their residual depths as inputs and the corresponding expected costs as output. To digitise molecules into mathematical models, extended-connectivity fingerprint (ECFP) was applied, which is a topological fingerprint to convert the circular structure of neighborhood of each non-hydrogen atom into bytes. In this work, the radius of the fingerprint was four (ECFP4), which detects the multiple layers of the neighborhoods from the molecule centre, and all molecules were converted into 2048 fixed-length bit string. Overall, the input has 2049 features, from which 2048 are binary variables from ECFP, and one from the residual depth. |
61e53beceab6ef20aae41719 | 27 | Multi-layered perception (MLP) neural network was used as the machine learning model to learn from the data, and this was conducted by using the deep learning API Keras. Although over one million datapoints were obtained from each iteration, the structure of the MLP was simple to avoid data overfitting, especially from the 2048 binary variables. The MLP includes an input layer of 2049 nodes, followed by a batch normalization layer to standardize the inputs. |
61e53beceab6ef20aae41719 | 28 | Three hidden layers of 30, 15 and 5 nodes respectively using the 'exponential linear unit' activation function were added, and right after each hidden layer, three dropout layers, with dropout rate of 0.3, 0.2 and 0.1 were added to randomly reduce the size of hidden nodes to avoid overfitting. This was eventually followed by an output layer of one node, also with 'elu' activation function, which approximates the molecular expected cost. For specification, MLP used learning rate of 0.002 to slowly learn from data, 'mean square error' as the loss function, |
61e53beceab6ef20aae41719 | 29 | The logic of the policy model 'decision maker' is as follows: it determines costs of molecules based on their functional structures and previous synthetic performance and minimises the costs of synthetic planning by selecting the overall low-cost molecules. The expected costs of the molecules were learned through machine learning. Generally at each iteration, after 50 epochs of learning, the test data outputs usually show approximately 45% scaled root mean square error (RMSE) and 65% Pearson correlation coefficient (correlation) from the test data approximations. The RMSE and correlation equations and results for all three environments and all 20 iterations are shown in SI. The RMSEs are high since we tried to learn from 2049 features out of millions of molecules, and by no means could an MLP with three hidden layers fit all the costs of molecular structures using such a simple model structure. Also, we did not expect the MLP to grasp all details from the observations, since a large portion were from trialand-error noise, which would eventually cause overfitting. However, approximately 70% correlation means that the model learned the overall relationship among molecular structure, retrosynthetic depth and the expected costs, which was promising for overall predictions. |
61e53beceab6ef20aae41719 | 30 | With well-trained policy models, the optimisation results improved over the iterations. The statistics of the expected costs of molecules from the biological reactions pool over the 20 iterations is shown in Figure , and the chemical and hybrid reaction pathways show similar optimisation trends (Figure ). At iteration 0, the median of expected costs for all target molecules reaches approximately 100, which means in most cases, the random sampling 'decision maker' picks dead-end molecules to build reaction routes for the target molecules. |
61e53beceab6ef20aae41719 | 31 | For a great portion of the outliers, the 'decision maker' selects multiple dead-end molecules, which approaches the expected costs of multiple hundreds. By learning from trial-and-error results, the policy model reduces the expected costs of most target molecules, with median expected costs being stabilised below 10 in the last five iterations and finalised at 5.2 at the last iteration. Along the 20 iterations, although the portion of outliers also reduces, there are still outliers that reach costs over 200 in the last five iterations. These are large protein molecules which usually have molar weights over 500 and are believed to be hard to synthesize, which include C16-KDO2-lipid A, UDP-4-amino-4-deoxy-L-arabinose, etc. The situation of target molecule ferricytochrome c has not been improved over the 20 iterations, which stabilises at the expected costs of 704 in the biological reaction pathway in Figure . |
61e53beceab6ef20aae41719 | 32 | The expected costs of target molecules rely on the molecule synthetic difficulty and the quality of the decision maker to reach a near-optimal synthetic route. Since we used large target molecules pools and fix the molecules synthetic difficulty by filtering the molecules SMILES string length from 20 to 100, the synthetic difficulty was fair for the chemical, biological and hybrid reactions pools. Hence, we use the median expected costs of target molecules to judge the optimisation results from the three reactions pools. As the policy model 'decision makers' being trained and optimised, whilst the median expected cost of the target molecules from the biological reactions pool has a significant jump at iteration 12, those from the chemical and the hybrid reactions pools both reduce gradually over the 20 iterations (Figure ). The three curves all tend to be stable in the last five iterations, which means they all approach optimisation limits. |
61e53beceab6ef20aae41719 | 33 | This can be interpreted such that in most cases, the molecule synthetic difficulty reduces in the hybrid reactions pool compared with the organic synthesis or synthetic biology ones alone. It also suggests that although the addition of the biological dataset only adds 0.36% data into the chemical dataset (Figure in terms of the number of reactions), in overall, it adds value by 3.4% to the organic synthesis to reach better synthetic results (by comparing the expected costs of molecule medians of 4.3 and 4.15 in organic and hybrid synthesis respectively): it is able to improve the redox efficiency and find more opportunities for synthetic shortcuts among molecules via hybridising the reactions pools. |
61e53beceab6ef20aae41719 | 34 | However, we acknowledge that this interpretation is specific to the used assessment criteria and penalty scores. Other advantages of biological reactions such as greenness and close-to-ambient reaction conditions have not been covered by the current methodology. We also did not implement any quantification of the drawbacks of biological reactions. For example, it is common for biological reactions to be highly dependent on the rest of cellular metabolic network, which increases operational costs of reactions. We also did not consider product separation for any of the reactions in the current implementation. |
61e53beceab6ef20aae41719 | 35 | KEGG drugs database gives a list of drug molecules as active pharmaceutical ingredients. To test the performance of the final policy model 'decision makers', molecules were crawled from the database to compute reaction pathways to them. Since the reaction pathways were compared by reactions pools from organic chemical, biological and hybrid databases, only 3,821 drug molecules co-existing in the chemical and biological datasets were used as the target drug molecules to compare the optimal reaction pathways from the three datasets. Here we also filtered these molecules into a set of 3,746, to contain only molecules with SMILES string length greater than 10 to increase the synthetic planning complexity, see SI. The final policy model 'decision makers' from the three reactions pools were used to identify the synthetic pathways for each of the drug molecules. The results from the three environments are shown in Figure . Different from other random selected target molecules with shorter SMILES string lengths, the drug molecules are more difficult to find synthetic routes. Whilst the target molecules are usually being synthesised within five steps, the cost to make drug molecules reach median of 100 for the three datasets, which means the routes always point to a dead-end molecule. This indicates that due to the molecular complexity, the majority of drug molecules cannot be synthesised with the demonstrated method and datasets. For further work, we could use partial reactions to predict functional transformations. In this way, more possible solutions could be given to the synthetic routes. However, it is also seen that the hybrid environment exhibits a heavier tail towards lower costs to make the molecules. This means the method opens possibilities to synthesise a significant proportion of drug molecules making use of the full set of chemistry combining organic synthesis and synthetic biology. |
61e53beceab6ef20aae41719 | 36 | An example of these successfully synthesised molecules is glucosinolate, an active pharmaceutical ingredient of multiple Chinese medicines, which are antibacterial, antioxidant, anticarcinogenic, etc. We illustrate the following 7-step synthetic route of glucosinolate in Scheme 1, suggested by the policy model. The cost of making glucosinolate in this route is 3.66, with five building block used. The depth of the longest branch is 7 steps. The route uses four organic chemical reactions and four synthetic biological reactions. Excluding free metabolites and cofactors such as NADPH, NADP+, oxygen, etc., the route produces in total three side products -UDP, pyruvate and carbapen-2-em-3-carboxylate; adenosine-3', 5'biphosphate from the first biological reaction (RxnBio0) is a coenzyme circulating over cell organisms, and thus, it is not strictly a side product. To compare, there is no chemical route to synthesise this molecule, and the cost of the purely biological route is 7.15. |
61e53beceab6ef20aae41719 | 37 | We presented an efficient method to suggest near-optimal biochemical pathway via data mining from organic chemistry and synthetic biology datasets, and reinforcement learning decision making. With this method, we also proved that, in overall, hybridising chemical and biological reactions to plan synthetic pathways was better than conventional organic synthesis by 3.4% with respect of the synthesis of all target molecules in the molecular space, due to advantages of synthetic biology to improve redox efficiency and enable synthetic shortcuts for the reaction routes. This conclusion was justified by using atom economy, numbers of reactions steps, and price of building blocks as key criteria to quantify retrosynthesis performance, and applying a comprehensive list of target molecules for the comparison of the organic chemical, synthetic biological and hybrid reactions pools. We could especially benefit from the well-trained policy model to plan the synthetic routes of a set of drug molecules. The case study of certain drug molecules indicates synthesis would be significantly eased with the help of synthetic biology reactions. This methodology could be further extended to mine more comprehensive reaction data to further understand the true costs of making biological reactions, which would make it possible to plan reaction routes with more confidence. |
654889ffc573f893f1e64546 | 0 | Part of the solution may come from the production of fuel and chemicals from lignocellulose which comprise an abundant and cheap resource that does not compromise food security. Yet, the cost-efficient production of biofuels and materials or chemicals from lignocellulose is currently hampered by its natural recalcitrance. Consequently, the discovery of a new class of enzymes that boosts the degradation of polysaccharides has attracted considerable attention. This enzyme class is now termed lytic polysaccharide monooxygenases (LPMOs) and is currently grouped into eight families: AA9 to AA11 and AA13 to AA17. The LPMOs differ in substrate-specificity but they generally catalyze the oxidation of the glycosidic C-H bonds in the polysaccharide, leading ultimately to a cleavage of the glycosidic bond. This cleavage causes the boosting effect and is administered by a copper-containing active site with Cu coordinated by two histidines that has been coined the histidine brace. The active site with the histidine brace in the Cu(II)-resting state is shown in Fig. . |
654889ffc573f893f1e64546 | 1 | The oxidation of inactive C-H bonds requires a potent oxidant; several theoretical studies have shown that a Cu(II)-oxyl is sufficiently potent. This Cu(II)-oxyl can be formed with a suitable co-substrate after reduction of the Cu(II)-resting state (see Fig. , structure 1). Whether this co-substrate is O 2 or H 2 O 2 is still an ongoing debate, although some LPMOs have been shown to use exclusively hydrogen peroxide. While a major focus in LPMO research has been the investigation of the substrate oxidation mechanism, the potent oxidative species can also cause oxidative damage to LPMOs in the absence of a substrate. This self-oxidation remains poorly understood but renders LPMOs inactive and thus hinders efficient exploitation of the enzymes. Proteomics Figure : a) Formation of Cu(II)-oxyl 1 from the Cu(II)-resting state (in gray) from a reductant and hydrogen peroxide, and generation of intermediates 2 and 3 from 1. Intermediates 1-3 (in black) are investigated here for their role as part of oxidative damage or protective pathways (species in gray are not included in this study). b) Ball-and-stick model of the first coordination sphere of the QM/MM optimized Cu(II)-resting state of LsAA9 employing TPSS/def2-SV(P). |
654889ffc573f893f1e64546 | 2 | techniques showed that oxidative damage is primarily confined to the two histidine residues coordinating the copper. We recently employed QM/MM to show that the oxidation of the histidine brace can be initiated by the hydrogen abstraction from either of the two histidines by a Cu(II)-oxyl (1), forming a species (3) with a histidyl radical coupled to a |
654889ffc573f893f1e64546 | 3 | [CuOH] + moiety (see reaction II in Fig. ). In a complementary series of experimental and theoretical investigations, we and others have proposed that a tyrosine close to the active site effectively protects LPMOs against the self-oxidation processes by forming a tyrosyl radical (2). The tyrosyl radical (2) can be formed from by reaction I in Fig. where the Cu(II)oxyl species (1) abstracts hydrogen from the hydroxo group of the tyrosine, 40 thus forming a |
654889ffc573f893f1e64546 | 4 | [CuOH] + moiety coordinated to a tyrosyl radical. Tyrosyl radicals have now been observed for a number of LPMOs and are characterized by having UV-vis spectra with strong, sharp absorption features around 400-420 nm. We note in passing that one investigation 39 (for LsAA9) showed rather different features, and based on quantum-mechanics/molecular mechanics and time-dependent density functional theory (TD-DFT), we 40 have argued that this spectrum originated in a different species (trans-[TyrO • -CuOH] + in Fig. ). |
654889ffc573f893f1e64546 | 5 | The tyrosine residue has long mystified LPMO researchers: it is widely conserved in most LPMO families, except for the majority of AA10 LPMOs, where it is replaced by phenylalanine. It seems now that it's role is to initialize a hole-hopping mechanism, known from other oxoreductase enzymes: a radical species ("hole") is generated close to the active site by a highly oxidizing species; this hole is then directed away (towards the surface) through electron-transfer chains comprised of aromatic/redox-active amino acids. Yet, detailed investigations of these reactions have been rare for oxoreductases; 44,45 and has only recently begun to appear for LPMOs. To confuse matter more, a hitherto undetected histidyl species was recently claimed to be observed through transient spectroscopy in LsAA9. The histidyl radical was assigned based on UV-vis absorption at 360 nm and high-energy-resolution fluorescence-detected X-ray absorption spectroscopy (HERFD-XAS). |
654889ffc573f893f1e64546 | 6 | Several open questions now remain concerning the initial parts of the hole-hopping mechanism: (i) How does the tyrosine radical formation compare energetically, relative to the histidyl radical formation? (ii) Are both tyrosyl and histidyl radicals part of the protective mechanism? (iii) Do different LPMOs display different mechanisms? The last open question arises since previous QM/MM calculations on the protective or oxidative damage mechanisms focused exclusively on a particular member of the AA9 family, namely LsAA9, whereas experimental studies focused on different LPMOs In this paper we will address (i) and (ii) by employing QM/MM calculations for reactions I-III in Fig. . We will also address (iii) by, for the first time, directly comparing oxidative damage and protective mechanisms of two enzymes, namely LsAA9 and TaAA9 (see Fig. ). One of our findings is that LsAA9 and TaAA9 (that have quite similar second coordination spheres) occasionally display significant differences in their mechanisms. Given that little is known concerning the initial events of these protective mechanisms, our findings are likely to be highly relevant for other oxoreductases as well. |
654889ffc573f893f1e64546 | 7 | QM/MM calculations: All QM/MM calculations were based on equilibrated and QM/MM optimized structures of LsAA9 from Lentinus similis (5ACF ) and TaAA9A from Thermoascus aurantiacus (2YET ) from the studies ref. The second-sphere tyrosine is present in both enzymes (Tyr175 and Tyr164 in TaAA9 and LsAA9, respectively, as shown in Fig. ) and in both LPMOs a tyrosyl species was spectroscopically characterised. The protonation state of the second-coordination sphere histidine (His164) in TaAA9 was changed from a doubly protonated form (Hip164) to a singly protonated one (Hie164), to match the protonation state present in the LsAA9 structure. The employed QM regions are shown in Fig. . For LsAA9 the QM region comprised the copper ion, the oxyl/hydroxyl ligand, the complete methylated His1 residue, and the side chains of His78, Tyr164, Thr2, Gln162, Glu148 and Hie147 as well as four water molecules. This is the same QM region also employed in ref. and a slightly larger QM region (extended by Glu143 and three water molecules) as employed in ref. . For TaAA9, the QM region consists of the copper ion and all the residues of the first coordination sphere, i.e., the full His1 residue, the oxyl/hydroxyl ligand, the imidazole ring of His86, the phenol ring of Tyr175 and the axial water molecule. Additionally, the side chains of Gln173 and Hie164 were included, as well as parts of the backbone from the Gly2 residue. The latter was included up to the C α atom and capped by replacing the carboxyl C with a hydrogen atom. The same applies to the corresponding residue Thr2 in LsAA9. The side chains of all other residues were capped with a hydrogen by replacing C α . Note that TaAA9 does not have a Glu residue corresponding to Glu148 in LsAA9, and we therefore did not include the residue at this position in the QM region for TaAA9. |
654889ffc573f893f1e64546 | 8 | All structure optimizations and energy calculations employed a substractive QM/MM approach with electrostatic embedding as implemented in the modular program ComQum. ComQum interfaces the QM software Turbomole 51 and the MM software AMBER. For all the structure optimizations the dispersion-corrected TPSS-D3 functional with Becke-Johnson damping and a def2-SV(P) basis set were employed. For all calculations with TPSS the resolution of identity (RI) approximation with standard auxiliary basis sets was applied. All atoms in the MM region were kept fixed during the geometry optimizations. |
654889ffc573f893f1e64546 | 9 | The reaction and activation energies were computed as linear transit calculations without thermochemical or zero-point vibrational energy corrections as these have shown to be small for hydrogen abstractions. The reported energies for the reactant, product and transition state structures were obtained from single-point calculations on the QM/MM optimized structures, employing the def2-TZVPP basis set (with an auxiliary basis set of the same size) and the functionals TPSS-D3 and B3LYP-D3. The single-point calculations include protein electrostatics, and a MM contribution calculated with TPSS/def2-SV(P). For brevity, we generally denote these functionals as TPSS and B3LYP throughout the paper although all calculations (for energies and geometries) always included dispersion corrections. On several occasions we compare to previous calculations, where generally TPSS/def2-SV(P) was used for geometry optimizations and B3LYP/def2-TZVPD for single point energies. |
654889ffc573f893f1e64546 | 10 | Our calculations generally started from the [CuO] + species (1 in Fig. ). This intermediate was selected as previous calculations have shown that this is the most likely species to abstract hydrogen from the substrate, and it is readily formed from the reaction between the reduced resting state and H 2 O 2 . From these previous calculations, we know that this [CuO] + moiety is most accurately described as a Cu 2+ and an O -• radical, spin-coupled to either a triplet or an open-shell singlet. Thus, the intermediates considered here are generally expected to attain either singlet and triplet spin states, and we have always attempted to obtain both of these spin states. As previous studies showed the closed-shell singlet states to be significantly higher in energy, the singlet states were only calculated as open-shell species in this study. The open-shell singlets were obtained in a spin-unrestricted (broken symmetry) formulation and the calculations were typically initiated from the triplet-state structures (using the triplet molecular orbital coefficients as initial guess). Similarly, for calculations with def2-TZVPP the open-shell singlet calculations were always started from the triplet molecular orbital coefficients. The convergence to an open-shell singlet was always confirmed by inspection of the Mulliken spin densities (which are reported in the supporting information). We compare the obtained electronic structure of 1 to previous calculations involving the Cu(II)-oxyl (1) intermediate. However, since this has been discussed frequently in the literature, this discussion is moved to the SI. |
654889ffc573f893f1e64546 | 11 | For reaction II of TaAA9 we were unable to locate an open-shell singlet for 3, although it could be obtained for a conformer of 3, which we here denote 3 ′ . The calculations along reaction II collapse into a closed-shell singlet along the linear transit, the last open-shell singlet energy being 125 kJ/mol (distance between O ox and H OH-Tyr restraint to 1.27 Å) for B3LYP and 84 kJ/mol (distrance restraint to 1.30 Å) for TPSS, cf. Table . This issue has also been observed previously for the TPSS functional with LsAA9. In these cases, we base the discussion on the triplet state. |
654889ffc573f893f1e64546 | 12 | For the calculations of reaction III, we observed for both LsAA9 and TaAA9 that the QM/MM energies for the barrier showed high MM contributions (>33 kJ/mol). We analyzed the energy contributions from individual residues and observed that this was caused by a residue close to the tyrosine (Phe43 and Pro30 in LsAA9 and TaAA9, respectively). Hence, we chose to include these residues in the QM region for this reaction, reducing the MM contributions to ∼13 kJ/mol for both enzymes. Since the reaction for TaAA9 was more favorable for the triplet state and no open-shell singlet could be located for LsAA9, we base the discussion on the triplet state. |
654889ffc573f893f1e64546 | 13 | UV-vis spectra UV-vis spectra were calculated for intermediates 2 and 3/3 ′ considering both open-shell singlet and triplet spin states, if open-shell singlet states were obtained. We performed TD-DFT calculations in Gaussian 16, employing the CAM-B3LYP functional and def2-TZVPP basis set (both as implemented in Gaussian) and including 45 states (roots). |
654889ffc573f893f1e64546 | 14 | For better comparison to the spectra calculated for LsAA9 in ref., 40 the input structures for LsAA9 and TaAA9 in the TD-DFT calculations were slightly truncated compared to the QM system that was employed for the structure optimizations and energy calculations. For LsAA9, Thr2 was removed to the amide N, and for TaAA9 only the amide-group of residue Gly2 was included. The cut bonds were saturated with a hydrogen atom (see Fig. in the SI). Calculated oscillator strengths and energies were convoluted using a Gaussian function and a broadening factor of 0.3 eV. To prope the electrostatic effect of the environment, we in one case (for 2 of TaAA9 in the triplet state) added point-charges from the QM/MM calculations. |
654889ffc573f893f1e64546 | 15 | We start by comparing LsAA9 and TaAA9 for the tyrosyl radical (2) formation directly from the Cu(II)-oxyl ( ) intermediate (i.e., reaction I in Fig. ). Next, we compare these two AA9 LPMOs for histidyl radical (3) formation (i.e. reaction II in Fig. ). We finally discuss whether intermediates 2 and 3 can be inter-converted (reaction III in Fig. ) as recently suggested. 41 |
654889ffc573f893f1e64546 | 16 | While there are some differences between the two employed functionals, both agree that the activation energies are consistently lower for TaAA9: we obtain a barrier of 44 kJ/mol (28 kJ/mol with TPSS), compared to a reaction barrier of 77 kJ/mol for LsAA9 (38 kJ/mol with TPSS). Both functionals also agree that the reaction is exothermic. In fact, the final reaction energies are quite similar between the two proteins, where we obtain an energy of -35 kJ/mol for TaAA9 (-76 kJ/mol with TPSS) and -44 kJ/mol for LsAA9 (-75 kJ/mol with TPSS). The result is qualitatively similar to previous calculations for LsAA9 where a smaller QM region was employed; 40 here the reaction barrier for reaction I was predicted to be 64 kJ/mol (53 kJ/mol with TPSS). We can thus conclude that LsAA9 and TaAA9 are likely to form the same tyrosyl intermediate, albeit with different kinetics. . |
654889ffc573f893f1e64546 | 17 | We can also compare structures of 1 and 2 as well as the transition state in the two LP-MOs in Fig. . For both enzymes, the largest structural change between 1 and 2 is the change in the Cu-O Tyr bond, which shortens from 2.6 Å (both enzymes) in 1 to 2.4 Å (TaAA9) or 2.2 Å (LsAA9) in 2. Thus, the de-protonation of tyrosine leads to coordination of the tyrosyl, consistent with a previous calculation on LsAA9. In both enzymes, the tyrosine OH-group forms a hydrogen bond to a Gln residue in 1 (the bond-distance is 1.6 Å in both TaAA9 and LsAA9). This hydrogen bond is partly broken in the transition state; here TaAA9 and LsAA9 are different as the bond distance is 2.3 Å in TaAA9 and 2.1 Å in LsAA9. |
654889ffc573f893f1e64546 | 18 | Notably, the larger distance in TaAA9 means that we can optimize a stable intermediate after breaking of the hydrogen bond to Gln, but this intermediate is close to degenerate with TS I , and we have therefore not included it in Fig. (a full Figure including this intermediate is given in the SI, see Fig. and). Interestingly, the hydrogen bond to Gln is re-formed in 2 (with a distance of 1.7 Å in TaAA9 and 1.8 Å in LsAA9), where the OH group is now coordinated to Cu(II). Other distances are roughly similar (cf. Table ), so the energetic differences between the reaction barrier of the two LPMOs in Fig. may be traced back to the structural differences in the local hydrogen bonding of Gln to the transition state. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.