Conspectus Photochemical reactions are of great importance in chemistry, biology, and materials science because they take advantage of a renewable energy source, mild reaction conditions, and high ...atom economy. Light absorption can excite molecules to a higher energy electronic state of the same spin multiplicity. The following nonadiabatic processes induce molecular transformations that afford exotic molecular architectures and high-energy-isomers that are inaccessible by thermal means. Computational simulations now complement time-resolved instrumentation to reveal ultrafast excited-state mechanistic information for photochemical reactions that is essential in disentangling elusive spectroscopic features, excited-state lifetimes, and excited-state mechanistic critical points. Nonadiabatic molecular dynamics (NAMD), powered by surface hopping techniques, is among the most widely applied techniques to model the photochemical reactions of medium-sized molecules. However, the computational efficiency is limited because of the requisite thousands of multiconfigurational quantum-chemical calculations multiplied by hundreds of trajectories. Machine learning (ML) has emerged as a revolutionary force in computational chemistry to predict the outcome of the resource-intensive multiconfigurational calculations on the fly. An ML potential trained with a substantial set of quantum-chemical calculations can predict the energies and forces with errors under chemical accuracy at a negligible cost. The integration of ML potentials in NAMD dramatically extends the maximum simulation time scale by ∼10 000-fold to the nanosecond regime. In this Account, we present a comprehensive demonstration of ML photodynamics simulations and summarize our most recent applications in resolving complex photochemical reactions. First, we address three fundamental components of ML techniques for photodynamics simulations: the quantum-chemical data set, the ML potential, and NAMD. Second, we describe best practices in building training data and our procedure toward training the ML photodynamics model with our recent literature contributions. We introduce a convenient training data generation scheme combining Wigner sampling and geometrical interpolation. It trains reliable and effective ML potentials suitable for subsequent active learning to detect undersampled data. We demonstrate how active learning automatically discovers new mechanistic pathways and reproduces experimental results. We point out that atomic permutation is an essential data augmentation approach to improve the learnability of distance-based molecular descriptors for highly symmetric molecules. Third, we demonstrate the utility of ML-photodynamics by showing the results of ML photodynamics simulations of (1) photo-torquoselective 4π disrotatory electrocyclic ring closing of norbornyl cyclohexadiene, which reveals a thermal conversion from experimentally unobserved intermediates to the reactant in 1 ns; (2) 2 + 2 photocycloaddition of substituted 3-syn-ladderdienes in competition with 4π and 6π electrocyclic ring-opening reactions, uncovering substituent effects to explain the reported increased quantum yield of substituted cubane precursors; and (3) photochemical 4π disrotatory electrocyclic reactions of fluorobenzenes in nanoseconds with XMS-CASPT2-level training data. We expect this Account to broaden understanding of ML photodynamics and inspire future developments and applications to increasingly large molecules within complex environments on long time scales.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
Benzene fluorination increases chemoselectivities for Dewar‐benzenes via 4π‐disrotatory electrocyclization. However, the origin of the chemo‐ and regioselectivities of fluorobenzenes remains ...unexplained because of the experimental limitations in resolving the excited‐state structures on ultrafast timescales. The computational cost of multiconfigurational nonadiabatic molecular dynamics simulations is also currently cost‐prohibitive. We now provide high‐fidelity structural information and reaction outcome predictions with machine‐learning‐accelerated photodynamics simulations of a series of fluorobenzenes, C6F6‐nHn, n=0–3, to study their S1→S0 decay in 4 ns. We trained neural networks with XMS‐CASPT2(6,7)/aug‐cc‐pVDZ calculations, which reproduced the S1 absorption features with mean absolute errors of 0.04 eV (<2 nm). The predicted nonradiative decay constants for C6F4H2, C6F6, C6F3H3, and C6F5H are 116, 60, 28, and 12 ps, respectively, in broad qualitative agreement with the experiments. Our calculations show that a pseudo Jahn–Teller distortion of fluorinated benzenes leads to an S1 local‐minimum region that extends the excited‐state lifetimes of fluorobenzenes. The pseudo Jahn–Teller distortions reduce when fluorination decreases. Our analysis of the S1 dynamics shows that the pseudo‐Jahn–Teller distortions promote an excited‐state cis‐trans isomerization of a πC‐C bond. We characterized the surface hopping points from our NAMD simulations and identified instantaneous nuclear momentum as a factor that promotes the electrocyclizations.
Machine learning accelerated photodynamics simulations reveal the excited‐state mechanistic pathways of fluorobenzenes. The excited‐state distortions promote the cis‐trans isomerization of a πC‐C bond during the nonradiative decay. The distribution of surface hopping structures indicates the photochemical 4π‐electrocyclizations of fluorobenzene originate from the dynamical effects, where the instantaneous nuclear momentum drives the σ‐bond formation in the Dewar‐fluorobenzenes.
Full text
Available for:
BFBNIB, FZAB, GIS, IJS, KILJ, NLZOH, NUK, OILJ, SBCE, SBMB, UL, UM, UPUK
The hippocampus is essential for different forms of declarative memory, including social memory, the ability to recognize and remember a conspecific. Although recent studies identify the importance ...of the dorsal CA2 region of the hippocampus in social memory storage, little is known about its sources of social information. Because CA2, like other hippocampal regions, receives its major source of spatial and non-spatial information from the medial and lateral subdivisions of entorhinal cortex (MEC and LEC), respectively, we investigated the importance of these inputs for social memory. Whereas MEC inputs to CA2 are dispensable, the direct inputs to CA2 from LEC are both selectively activated during social exploration and required for social memory. This selective behavioral role of LEC is reflected in the stronger excitatory drive it provides to CA2 compared with MEC. Thus, a direct LEC → CA2 circuit is tuned to convey social information that is critical for social memory.
•Lateral entorhinal cortex (LEC) inputs to hippocampal CA2 underlie social memory•Medial entorhinal cortex (MEC) CA2 input is weak and not involved in social memory•Social memory requires the direct but not indirect LEC inputs to CA2•LEC CA2 inputs are selectively activated by social over non-social exploration
Although the CA2 hippocampal region is essential for social memory and detection of social novelty, the inputs that provide social information to CA2 are unknown. We found that social memory depends on the direct inputs CA2 receives from the lateral entorhinal cortex. As this input is activated similarly by novel and familiar individuals, CA2 itself may compute novelty.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Photochemical 2 + 2-cycloadditions store solar energy in chemical bonds and efficiently access strained organic molecular architectures. Functionalized 3-ladderdienes undergo 2 + ...2-photocycloadditions to afford cubanes, a class of strained organic molecules. The substituents (e.g., methyl, trifluoromethyl, and cyclopropyl) affect the overall reactivities of these cubane precursors; the yields range from 1 to 48%. However, the origin of these substituent effects on the reactivities and chemoselectivities is not understood. We now integrate single and multireference calculations and machine-learning-accelerated nonadiabatic molecular dynamics (ML-NAMD) to understand how substituents affect the ultrafast dynamics and mechanism of 2 + 2-photocycloadditions. Steric clashes between substituent groups destabilize the 4π-electrocyclic ring-opening pathway and minimum energy conical intersections by 0.72–1.15 eV and reaction energies by 0.68–2.34 eV. Noncovalent dispersive interactions stabilize the 2 + 2-photocycloaddition pathway; the conical intersection energies are lower by 0.31–0.85 eV, and the reaction energies are lower by 0.03–0.82 eV. The 2 ps ML-NAMD trajectories reveal that closed-shell repulsions block a 6π-conrotatory electrocyclic ring-opening pathway with increasing steric bulk. Thirty-eight percent of the methyl-substituted 3-ladderdiene trajectories proceed through the 6π-conrotatory electrocyclic ring-opening, whereas the trifluoromethyl- and cyclopropyl-substituted 3-ladderdienes prefer the 2 + 2-photocycloaddition pathways. The predicted cubane yields (H: 0.4% < CH3: 1% < CF3: 14% < cPr: 15%) match the experimental trend; these substituents predistort the reactants to resemble the conical intersection geometries leading to cubanes.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
Sunlight is a renewable energy source that can be stored in chemical bonds using photochemical reactions. The synthesis of exotic and strained molecules is especially attractive with photochemical ...techniques because of the associated efficient and mild reaction conditions. We have understood the photophysics and subsequent photochemistry of a possible cubane precursor, tricyclo4,2,0,02,5octa-3,7-diene with complete active space self-consistent field (CASSCF) calculations with an (8,7) active space and the ANO-S-VDZP basis set to. The CASSCF energies were corrected with a second-order perturbative correction CASPT2(8,7)/ANO-S-VDZP. The S0 → S1 vertical excitation energy of 1 is 6.25 eV, which is a π → π* excitation. The minimum energy path from the S1 Franck–Condon point leads to a 4π-disrotatory electrocyclic ring-opening reaction to afford bicyclo4,2,0octa-2,4,7-triene. The 2D potential energy surface scan located a rhomboidal S1/S0 minimum energy crossing point connecting 1 and cubane, suggesting that a cycloaddition is theoretically possible. We used the fewest switches surface hopping to study the photodynamics of this cycloaddition: 85% of 1722 trajectories relaxed to eight products; the major products are bicyclo4,2,0octa-2,4,7-triene (30%) and cycloocta-1,3,5,7-tetraene (32%). Only 0.4% of trajectories undergo a 2 + 2 cycloaddition to form cubane.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
Photochemical reactions exemplify “green” chemistry and are an essential tool for synthesizing highly strained molecules under mild conditions with light. The light-promoted denitrogenation of ...bicyclic azoalkanes affords functionalized, stereoenriched bicyclo1.1.0butanes. These reactions were revisited with multireference calculations and non-adiabatic molecular dynamics (NAMD) simulations to provide a detailed analysis of the photophysics, reactivities, and unexplained stereoselectivity of a series of diazabicyclo2.1.1hexenes. We used complete active space self-consistent field (CASSCF) calculations with an (8,8) active space and ANO-S-VDZP basis set; the CASSCF energies were corrected with CASPT2 (8,8)/ANO-S-VDZP. The nature of the electronic excitation is n → π* and ranges from 3.77 to 3.91 eV for the diazabicyclo2.1.1hexenes reported here. Minimum energy path calculations showed stepwise C–N bond breaking and led directly to a minimum energy crossing point, corresponding to a stereochemical “double inversion” product. Wigner sampling of diazabicyclo2.1.1hexene provided initial conditions for 692 NAMD trajectories. We identified competing complete stereoselective and stereochemical scrambling pathways. The stereoselective pathways feature concerted bicyclobutane inversion and N2 extrusion. The stereochemical scrambling pathways involve N2 extrusion followed by bicyclobutane planarization, leading to stereochemical scrambling. The predicted diastereomeric excess (d.e.) almost exactly matches the experiment (calc.d.e. = 46% vs exp.d.e. = 47%). Our NAMD simulations with 672, 568, and 596 trajectories for 1-F, 1-Cl, and 1-Br predicted a d.e. of 94–97% for the double inversion products. Halogenation significantly perturbs the potential energy surface (PES) toward the retention products due to hyperconjugative interactions. The nC → σ*C–X, X = F, Cl, Br hyperconjugative effect leads to a broader shoulder region on the PES for double inversion.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
There are few studies comparing lesion evolution across different CNS demyelinating diseases, yet knowledge of this may be important for diagnosis and understanding differences in disease ...pathogenesis. We sought to compare MRI T2 lesion evolution in myelin oligodendrocyte glycoprotein immunoglobulin G (IgG)-associated disorder (MOGAD), aquaporin 4 IgG-positive neuromyelitis optica spectrum disorder (AQP4-IgG-NMOSD), and multiple sclerosis (MS).
In this descriptive study, we retrospectively identified Mayo Clinic patients with MOGAD, AQP4-IgG-NMOSD, or MS and (1) brain or myelitis attack; (2) available attack MRI within 6 weeks; and (3) follow-up MRI beyond 6 months without interval relapses in that region. Two neurologists identified the symptomatic or largest T2 lesion for each patient (index lesion). MRIs were then independently reviewed by 2 neuroradiologists blinded to diagnosis to determine resolution of T2 lesions by consensus. The index T2 lesion area was manually outlined acutely and at follow-up to assess variation in size.
We included 156 patients (MOGAD, 38; AQP4-IgG-NMOSD, 51; MS, 67) with 172 attacks (brain, 81; myelitis, 91). The age (median range) differed between MOGAD (25 2-74), AQP4-IgG-NMOSD (53 10-78), and MS (37 16-61) (
< 0.01) and female sex predominated in the AQP4-IgG-NMOSD (41/51 80%) and MS (51/67 76%) groups but not among those with MOGAD (17/38 45%). Complete resolution of the index T2 lesion was more frequent in MOGAD (brain, 13/18 72%; spine, 22/28 79%) than AQP4-IgG-NMOSD (brain, 3/21 14%; spine, 0/34 0%) and MS (brain, 7/42 17%; spine, 0/29 0%) (
< 0.001). Resolution of all T2 lesions occurred most often in MOGAD (brain, 7/18 39%; spine, 22/28 79%) than AQP4-IgG-NMOSD (brain, 2/21 10%; spine, 0/34 0%) and MS (brain, 2/42 5%; spine, 0/29 0%) (
< 0.01). There was a larger median (range) reduction in T2 lesion area in mm
on follow-up axial brain MRI with MOGAD (213 55-873) than AQP4-IgG-NMOSD (104 0.7-597) (
= 0.02) and MS (36 0-506) (
< 0.001) and the reductions in size on sagittal spine MRI follow-up in MOGAD (262 0-888) and AQP4-IgG-NMOSD (309 0-1885) were similar (
= 0.4) and greater than in MS (23 0-152) (
< 0.001).
The MRI T2 lesions in MOGAD resolve completely more often than in AQP4-IgG-NMOSD and MS. This has implications for diagnosis, monitoring disease activity, and clinical trial design, while also providing insight into pathogenesis of CNS demyelinating diseases.
Optical imaging in the shortwave infrared region (SWIR, 1000–2000 nm) provides high-resolution images in complex systems. Here we explore substituent placement on dimethylamino flavylium polymethine ...dyes, a class of SWIR fluorophores. We find that the position of the substituent significantly affects the λmax and fluorescence quantum yield. Quantum-mechanical calculations suggest that steric clashes control the extent of π-conjugation. These insights provide design principles for the development of fluorophores for enhanced SWIR imaging.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
The azide–dibenzocyclooctyne and trans-cyclooctene–tetrazine cycloadditions are both bioorthogonal and mutually orthogonal: trans-cyclooctene derivatives greatly prefer to react with tetrazines ...rather than azides, while dibenzocyclooctyne derivatives react with azides but not with tetrazines under physiological conditions. DFT calculations used to identify the origins of this extraordinary selectivity are reported, and design principles to guide discovery of new orthogonal cycloadditions are proposed. Two new bioorthogonal reagents, methylcyclopropene and 3,3,6,6-tetramethylthiacycloheptyne, are predicted to be mutually orthogonal in azide and tetrazine cycloadditions.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
The recent Zika virus (ZIKV) outbreak in Brazil has suggested associations of this virus infection with neurological disorders, including microcephaly in newborn infants and Guillian-Barré syndrome ...in adults. Previous reports have shown that AXL, a transmembrane receptor tyrosine kinase protein, is essential for ZIKV infection of mammalian cells, but this remains controversial. Here, we have assessed the involvement of AXL in the ability of ZIKV to infect mammalian cells, and also the requirement for endocytosis and acidic pH. We demonstrated that AXL is essential for ZIKV infection of human fibroblast cell line HT1080 as the targeted deletion of the gene for AXL in HT1080 cells made them no longer susceptible to ZIKV infection. Our results also showed that infection was prevented by lysosomotropic agents such as ammonium chloride, chloroquine and bafilomycin A1, which neutralize the normally acidic pH of endosomal compartments. Infection by ZIKV was also blocked by chlorpromazine, indicating a requirement for clathrin-mediated endocytosis. Taken together, our findings suggest that AXL most likely serves as an attachment factor for ZIKV on the cell surface, and that productive infection requires endocytosis and delivery of the virus to acidified intracellular compartments.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP