iPSC-derived organ-on-a-chip models for personalized human genetics and pharmacogenomics studies.

Genome-wide association studies (GWAS) have now correlated hundreds of genetic variants with complex genetic diseases and drug efficacy. Functional characterization of these factors remains challenging, particularly because of the lack of human model systems. Molecular and nanotechnological advances, in particular the ability to generate patient-specific PSC lines, differentiate them into diverse cell types, and seed and combine them on microfluidic chips, have led to the establishment of organ-on-a-chip (OoC) platforms that recapitulate organ biology. OoC technology thus provides unique personalized platforms for studying the effects of host genetics and environmental factors on organ physiology. In this review we describe the technology and provide examples of how OoCs may be used for disease modeling and pharmacogenetic research.

Covalent bond shortening and distortion induced by pressurization of thorium, uranium, and neptunium tetrakis aryloxides.

Covalency involving the 5f orbitals is regularly invoked to explain the reactivity, structure and spectroscopic properties of the actinides, but the ionic versus covalent nature of metal-ligand bonding in actinide complexes remains controversial. The tetrakis 2,6-di-tert-butylphenoxide complexes of Th, U and Np form an isostructural series of crystal structures containing approximately tetrahedral MO4 cores. We show that up to 3 GPa the Th and U crystal structures show negative linear compressibility as the OMO angles distort. At 3 GPa the angles snap back to their original values, reverting to a tetrahedral geometry with an abrupt shortening of the M-O distances by up to 0.1 Å. The Np complex shows similar but smaller effects, transforming above 2.4 GPa. Electronic structure calculations associate the M-O bond shortening with a change in covalency resulting from increased contributions to the M-O bonding by the metal 6d and 5f orbitals, the combination promoting MO4 flexibility at little cost in energy.

Exceptional uranium(VI)-nitride triple bond covalency from 15N nuclear magnetic resonance spectroscopy and quantum chemical analysis.

Determining the nature and extent of covalency of early actinide chemical bonding is a fundamentally important challenge. Recently, X-ray absorption, electron paramagnetic, and nuclear magnetic resonance spectroscopic studies have probed actinide-ligand covalency, largely confirming the paradigm of early actinide bonding varying from ionic to polarised-covalent, with this range sitting on the continuum between ionic lanthanide and more covalent d transition metal analogues. Here, we report measurement of the covalency of a terminal uranium(VI)-nitride by 15N nuclear magnetic resonance spectroscopy, and find an exceptional nitride chemical shift and chemical shift anisotropy. This redefines the 15N nuclear magnetic resonance spectroscopy parameter space, and experimentally confirms a prior computational prediction that the uranium(VI)-nitride triple bond is not only highly covalent, but, more so than d transition metal analogues. These results enable construction of general, predictive metal-ligand 15N chemical shift-bond order correlations, and reframe our understanding of actinide chemical bonding to guide future studies.

29Si NMR Spectroscopy as a Probe of s- and f-Block Metal(II)-Silanide Bond Covalency.

We report the use of 29Si NMR spectroscopy and DFT calculations combined to benchmark the covalency in the chemical bonding of s- and f-block metal-silicon bonds. The complexes [M(SitBu3)2(THF)2(THF)x] (1-M: M = Mg, Ca, Yb, x = 0; M = Sm, Eu, x = 1) and [M(SitBu2Me)2(THF)2(THF)x] (2-M: M = Mg, x = 0; M = Ca, Sm, Eu, Yb, x = 1) have been synthesized and characterized. DFT calculations and 29Si NMR spectroscopic analyses of 1-M and 2-M (M = Mg, Ca, Yb, No, the last in silico due to experimental unavailability) together with known {Si(SiMe3)3}--, {Si(SiMe2H)3}--, and {SiPh3}--substituted analogues provide 20 representative examples spanning five silanide ligands and four divalent metals, revealing that the metal-bound 29Si NMR isotropic chemical shifts, δSi, span a wide (∼225 ppm) range when the metal is kept constant, and direct, linear correlations are found between δSi and computed delocalization indices and quantum chemical topology interatomic exchange-correlation energies that are measures of bond covalency. The calculations reveal dominant s- and d-orbital character in the bonding of these silanide complexes, with no significant f-orbital contributions. The δSi is determined, relatively, by paramagnetic shielding for a given metal when the silanide is varied but by the spin-orbit shielding term when the metal is varied for a given ligand. The calculations suggest a covalency ordering of No(II) > Yb(II) > Ca(II) ≈ Mg(II), challenging the traditional view of late actinide chemical bonding being equivalent to that of the late lanthanides.

Polarised covalent thorium(IV)- and uranium(IV)-silicon bonds.

We report the synthesis and characterisation of isostructural thorium(iv)- and uranium(iv)-silanide actinide (An) complexes, providing an opportunity to directly compare Th-Si and U-Si chemical bonds. Quantum chemical calculations show significant and surprisingly similar An%:Si%, 7s-, 6d-, and 5f-orbital contributions from both elements in polarised covalent An-Si bonds, and marginally greater covalency in the U-Si vs. Th-Si linkages.

Quantum chemical topology and natural bond orbital analysis of M-O covalency in M(OC6H5)4 (M = Ti, Zr, Hf, Ce, Th, Pa, U, Np).

Covalency is complex yet central to our understanding of chemical bonding, particularly in the actinide series. Here we assess covalency in a series of isostructural d and f transition element compounds M(OC6H5)4 (M = Ti, Zr, Hf, Ce, Th, Pa, U, Np) using scalar relativistic hybrid density functional theory in conjunction with the Natural Bond Orbital (NBO), quantum theory of atoms in molecules (QTAIM) and interacting quantum atoms (IQA) approaches. The IQA exchange-correlation covalency metric is evaluated for the first time for actinides other than uranium, in order to assess its applicability in the 5f series. It is found to have excellent correlation with NBO and QTAIM covalency metrics, making it a promising addition to the computational toolkit for analysing metal-ligand bonding. Our range of metrics agree that the actinide-oxygen bonds are the most covalent of the elements studied, with those of the heavier group 4 elements the least. Within the early actinide series, Th stands apart from the other three elements considered, being consistently the least covalent.

Emergence of the structure-directing role of f-orbital overlap-driven covalency.

FEUDAL (f's essentially unaffected, d's accommodate ligands) is a longstanding bonding model in actinide chemistry, in which metal-ligand binding uses 6d-orbitals, with the 5f remaining non-bonding. The inverse-trans-influence (ITI) is a case where the model may break down, and it has been suggested that ionic and covalent effects work synergistically in the ITI. Here, we report an experimentally grounded computational study that quantitatively explores the ITI, and in particular the structure-directing role of f-orbital covalency. Strong donor ligands generate a cis-ligand-directing electrostatic potential (ESP) at the metal centre. When f-orbital participation, via overlap-driven covalency, becomes dominant via short actinide-element distances, this ionic ESP effect is overcome, favouring a trans-ligand-directed geometry. This study contradicts the accepted ITI paradigm in that here ionic and covalent effects work against each other, and suggests a clearly non-FEUDAL, structure-directing role for the f-orbitals.

Computational analysis of M-O covalency in M(OC6H5)4 (M = Ti, Zr, Hf, Ce, Th, U).

A series of compounds M(OC6H5)4 (M = Ti, Zr, Hf, Ce, Th, U) is studied with hybrid density functional theory, to assess M-O bond covalency. The series allows for the comparison of d and f element compounds that are structurally similar. Two well-established analysis methods are employed: Natural Bond Orbital and the Quantum Theory of Atoms in Molecules. A consistent pattern emerges; the U-O bond is the most covalent, followed by Ce-O and Th-O, with those involving the heavier transition metals the least so. The covalency of the Ti-O bond differs relative to Ce-O and Th-O, with the orbital-based method showing greater relative covalency for Ti than the electron density-based methods. The deformation energy of r(M-O) correlates with the d orbital contribution from the metal to the M-O bond, while no such correlation is found for the f orbital component. f orbital involvement in M-O bonding is an important component of covalency, facilitating orbital overlap and allowing for greater expansion of the electrons, thus lowering their kinetic energy.

Balancing Exchange Mixing in Density-Functional Approximations for Iron Porphyrin.

Predicting the correct ground-state multiplicity for iron(II) porphyrin, a high-spin quintet, remains a significant challenge for electronic-structure methods, including commonly employed density functionals. An even greater challenge for these methods is correctly predicting favorable binding of O2 to iron(II) porphyrin, due to the open-shell singlet character of the adduct. In this work, the performance of a modest set of contemporary density-functional approximations is assessed and the results interpreted using Bader delocalization indices. It is found that inclusion of greater proportions of Hartree-Fock exchange, in hybrid or range-separated hybrid functionals, has opposing effects; it improves the ability of the functional to identify the ground state but is detrimental to predicting favorable dioxygen binding. Because of the uncomplementary nature of these properties, accurate prediction of both the relative spin-state energies and the O2 binding enthalpy eludes conventional density-functional approximations.

Effect of amino acid ligands on the structure of iron porphyrins and their ability to bind oxygen.

Density functional theory is used to study a series of model iron porphyrins in the gas phase. In the first part of this study, three range-separated hybrid density functionals developed by Chai and Head-Gordon were assessed; ωB97, ωB97X, and ωB97XD. The effects of including full Hartree-Fock exchange at long-range and dispersion corrections are reported with respect to the geometries and binding energies of oxygen to the iron porphyrin systems. The functionals all correctly predict the quintet ground state for the deoxy-iron porphyrins, where typically hybrid functionals fail and predict a triplet ground state. Including dispersion in ωB97XD is shown to give the best results for the O2 binding energy and geometrical parameters. The second part of the study employs ωB97XD to study iron porphine systems with different amino acids in the axial position. Geometrical parameters are reported and compared to experimental data, where available. Binding energies of the systems with oxygen are also reported and discussed.

Self-assembling ADADA helices formed by hydrogen bonding.

A computational investigation of the electronic properties of an experimentally prepared ADADA helix indicates that the helix is held together with four strong hydrogen bonds as well as many other weak interactions. Determination of the electronic energy changes, as well as thermodynamic parameters, suggests that helix formation is a favorable process, driven by the formation of the hydrogen bonds. For instance, the unsubstituted helix has an electronic binding energy of -85.8 kJ/mol. Furthermore, the strength of binding can be tuned to some extent by the careful selection of substituents. The hydrogen bonds are strengthened when the pyridine ring (H-bond acceptor) is substituted with an electron-donating group such as an amine, while electron-withdrawing groups on the thiazine ring (H-bond donor) are preferred. The most significant enhancement in binding is achieved when the helix is constructed from monomers that consist of contiguous hydrogen-bond acceptors or donors. This so-called AAAAA-DDDDD helix exhibits a binding energy almost 4-fold greater than that of the unsubstituted ADADA helix at -335.4 kJ/mol, a dramatic improvement over the ADADA helix.

Ab initio investigation of the hydration of the tetrahedral perchlorate, perbromate, selenate, arsenate, and vanadate anions.

The geometries, energies, and vibrational frequencies of various isomers of XO(4)(m-)(H2O)n, x = Cl, Br, Se, As, V; n = 0-6, m = 1-3 are calculated at various levels up to MP2/6-31+G*. These properties are studied as a function of increasing cluster size. The experimental and theoretical vibrational spectra are compared where available.

Theoretical study of polaron formation in poly(G)-poly(C) cations.

Polaron formation in poly(G)-poly(C) cations is investigated with density functional theory (DFT) and molecular mechanics (MM) employing a two-layer ONIOM method. In these calculations, the high layer, composed of all complementary base pairs, is treated by a DFT method, while the low layer, which includes the sugar-phosphate backbone, counterions and water molecules, is described by the AMBER force field. The high layer is the model system in which the charge transfer takes place. According to our calculations, three or four guanines move in a paddle-like fashion when an electron is removed from the neutral model system. In the cation model system, about 80% of the charge is delocalized onto the guanine residues, and the remaining charge is delocalized onto the cytosine residues. This happens because guanine has a lower ionization potential (IP) than cytosine. The counterions and water molecules in the low layer are important in the geometry optimization. The optimized geometry of the model system is closer to the standard B-form structure when counterions and water molecules are included than when they are omitted. Comparison of the optimized neutral and cationic model systems reveals a polaron in poly(G)-poly(C) cations extending from the first to the third guanine. It is demonstrated that the position of counterions and the number of surrounding water molecules can affect polaron formation.

The effect of multiplicity on the size of iron(II) and the structure of iron(II) porphyrins.

The displacement of the iron(II) atom from the porphyrin plane in iron(II) porphyrin complexes is investigated with respect to the spin state of iron(II) employing density functional theory. In this study the quantum theory of atoms in molecules (QTAIM) is used to show that the atomic volume of iron is smaller in the quintet state of imidazolium ligated iron(II) porphyrin than in the triplet state. This is consistent with what has been found for free atoms and contradicts the original interpretation of structural studies with X-rays, which assumed that the out-of-plane displacement of iron from the porphyrin ring in the quintet state is due to the increased spatial size of the high-spin iron atom. The bonding environment of the iron atom is analyzed with respect to the electron density (ρ) at the bond critical points (BCPs). It is found that in the quintet state, relative to the triplet state, there is a stronger bonding interaction between iron and the nitrogen atoms of the porphyrin despite a longer bond length. It has previously been suggested that the weakening of these bonds is the cause of the out-of-plane displacement of iron. Since this is not the case, this implies that the magnitude of the bonding interaction between the iron atom and the axial ligand has a more significant role in the domed structure of the quintet state.

Synthesis of LJP 993, a multivalent conjugate of the N-terminal domain of beta2GPI and suppression of an anti-beta2GPI immune response.

LJP 993, a tetravalent conjugate of the amino-terminal domain (domain 1) of beta2GPI, was synthesized, and studies were carried out to explore the ability of LJP 993 to bind anti-beta2GPI antibodies and to function as a B cell toleragen. Domain 1 was expressed in Pichia pastoris, and the N-terminus was site-specifically modified by a transamination reaction converting the N-terminal glycine to a glyoxyl group. A tetravalent platform was synthesized with linkers that terminate in aminooxy groups. This was accomplished by preparing an ethylene glycol-based heterobifunctional linker that contains both a Boc-protected aminooxy group and a free primary amine. The linker was used to modify a tetravalent platform molecule by reacting the amino groups on the linker with 4-nitrophenyl carbonate esters on the platform to provide a linker-modified platform, and the Boc protecting groups were removed to provide a tetravalent aminooxy platform. Glyoxylated domain 1 was attached to the platform to provide LJP 993 by formation of oxime bonds. The protein domains of LJP 993 retain activity as evidenced by the ability of LJP 993 to bind to anti-beta2GPI antibodies. Dissociation constants (Kd) for domain 1 and LJP 993 bound to immobilized affinity-purified anti-beta2GPI antibodies from autoimmune thrombosis patients were determined using surface plasmon resonance. An immunized mouse model was developed to test the ability of LJP 993 to act as a toleragen. A thiol containing domain 1 analogue was expressed in insect cells using the baculovirus expression system, and it was used to prepare an immunogenic conjugate of domain 1 and maleimide-derivatized keyhole limpet hemocyanin (KLH). Mice were immunized with the KLH conjugate, and spleen cells were harvested from the immunized mice. The cells were incubated with various concentrations of LJP 993 and transferred to mice whose immune systems had been compromised by irradiation. The hosts were then boosted with the KLH-domain 1 conjugate, and after 7 days their antibody levels were measured. Host mice receiving cells that were treated with LJP 993 produced significantly lower amounts of anti-domain 1 antibodies than controls which received untreated cells, indicative of B cell tolerance.

Beta2-glycoprotein I-dependent anticardiolipin antibodies preferentially bind the amino terminal domain of beta2-glycoprotein I.

Many of the autoantibodies in antiphospholipid syndrome (APS) are directed against beta2-glycoprotein I (beta2-GPI). Recent studies from our laboratories have indicated that the immunodominant binding epitope(s) for high titer, affinity purified antibodies from 11 APS patients are localized to the amino terminal domain (domain 1) of beta2-GPI. The present study employed surface plasmon resonance to localize the immunodominant domain in serum samples from a large cohort of patients with GPL values ranging from 21 to 230 units (n = 106 patients). Eighty-eight percent of patients showed > or = threefold selectivity for beta2-GPI containing domain 1 relative to the domain deletion mutant that lacked domain 1. The domain 1 binding activity in patient serum was abolished by removing the IgG fraction from the serum and the binding activity could be fully reconstituted with the IgG fraction. Thus, analysis of serum samples from a large cohort of APS patients indicates that the immunodominant binding epitope(s) for anti-beta2 antibodies are localized to the amino terminal domain of beta2-GPI.

Multivalent thioether-peptide conjugates: B cell tolerance of an anti-peptide immune response.

Antibodies which bind beta2-glycoprotein I (beta2GPI) are associated with antiphospholipid syndrome. Synthetic peptide mimotopes have been discovered which compete with beta2GPI for binding to selected anti-beta2GPI. A thiol-containing linker was attached to the N-terminus of two cyclic thioether peptide mimotopes, peptides 1a and 1b. The resulting peptides, with linker attached, were reacted with two different haloacetylated platforms to prepare four tetravalent peptide-platform conjugates to be tested as B cell toleragens. The linker-containing peptides were reacted with maleimide-derivatized keyhole limpet hemocyanin (KLH) to provide peptide-KLH conjugates. Peptides 1a and 1b were also modified by acylation with 3-(4'-hydroxyphenyl)propionic acid N-hydroxysuccinimidyl ester. The resulting hydroxyphenyl peptides were radioiodinated and used to measure anti-peptide antibody levels. The KLH conjugates were used to immunize mice to generate an anti-peptide immune response. The immunized mice were treated with the conjugates or saline solution and boosted with the appropriate peptide-KLH conjugate. Three of the four conjugates suppressed the formation of anti-peptide antibody. The stabilities of the conjugates in mouse serum were measured, and the relative stabilities did not correlate with ability to suppress antibody formation.

Anti-beta2 glycoprotein I (beta2GPI) autoantibodies recognize an epitope on the first domain of beta2GPI.

Anticardiolipin (aCL) autoantibodies are associated with thrombosis, recurrent fetal loss, and thrombocytopenia. Only aCL found in autoimmune disease require the participation of the phospholipid binding plasma protein beta2 glycoprotein I (beta2GPI) for antibody binding and now are called anti-beta2GPI. The antigenic specificity of aCL affinity purified from 11 patients with high titers was evaluated in an effort to better understand the pathophysiology associated with aCL. Seven different recombinant domain-deleted mutants of human beta2GPI, and full length human beta2GPI (wild-type), were used in competition assays to inhibit the autoantibodies from binding to immobilized wild-type beta2GPI. Only those domain-deleted mutants that contained domain 1 inhibited the binding to immobilized wild-type beta2GPI from all of the patients. The domain-deleted mutants that contained domain 1 inhibited all aCL in a similar but not identical pattern, suggesting that these aCL recognize a similar, but distinguishable, epitope(s) present on domain 1.

A chemically defined, toleragen-based approach for targeting anti-beta2-glycoprotein I antibodies.

Antiphospholipid syndrome is characterized by a prothrombotic state and the presence of beta2-glycoprotein I (beta2-GPI)-dependent antiphospholipid antibodies. The feasibility of a B cell tolerance-based approach for specific reduction of anti-beta2-GPI antibodies was investigated. Anti-beta2-GPI antibodies isolated from a patient with antiphospholipid syndrome were used to screen peptide libraries expressed in phage, resulting in the identification of a phage that specifically bound anti-beta2-GPI antibodies. The phage-displayed peptide was identified and chemically optimized to generate a synthetic 14-mer peptide with an internal thioether linkage (LJP 685) that retained the binding profile of the original phage. LJP 685 was conjugated to a defined, non-immunogenic organic platform to generate a tetravalent presentation of LJP 685 for use as a toleragen. Tetravalent LJP 685 induced a dose-dependent reduction in antibody levels in mice previously immunized and boosted with LJP 685 coupled to the carrier keyhole limpet hemocyanin. These experiments support the technical feasibility of a tolerance-based approach for reducing anti-beta2-GPI antibodies in vivo.

Structural characterization and optimization of antibody-selected phage library mimotopes of an antigen associated with autoimmune recurrent thrombosis.

The presence of high titers of anti-cardiolipin antibodies (ACA's) of autoimmune origin, which are known to bind to plasma beta2-glycoprotein I (aka apolipoprotein H), correlates clinically with autoimmune recurrent thrombosis. Soluble beta2-glycoprotein I binds to solid-phase ACA (immobilized on a surface plasmon resonance chip) with a Kd of 1.4 microM, but if the reactants are reversed and beta2-glycoprotein I is on the solid-phase support, then the Kd is 52 nM. This 27-fold difference in affinity reflects the avidity/entropic advantage obtained for an antibody binding to an antigen that is made multivalent because it is attached to a solid phase. A mimotope of this antigen, selected from a phage display peptide library screen with an ACA, has been shown to bind to solid-phase ACA as a phage, using surface plasmon resonance. This peptide is representative of the motif from 37 peptides obtained in a previously reported phage library screen with this ACA (1). A synthetic version of this peptide, referred to as P4, has the sequence: A1G2P3C4I5L6L7A8R9D10R11C12P13G14, and binds to its selecting antibody with a Kd of 42 nM. NMR data indicate that proline-13 is present in both cis and trans configurations, and that these two geometries dramatically affect the overall tertiary structure of the molecule. The peptide lacking this proline binds severalfold better to the ACA, consistent with at least one of these structures having low affinity for binding ACA. Replacement of the arginine-9 position with a proline decreases binding affinity to ACA 10-fold. Another phage library-selected peptide has a proline in position 9, but also has a leucine in position 5, instead of isoleucine. Since its affinity for ACA is nearly as good as that for peptide P4, the phage library screening must have selected for a non-beta-branched amino acid in this position to compensate for the adverse effects of the arginine-9 to proline-9 substitution. The solution structure of a modified version of the antibody-selected phage peptide P4 with the central proline was determined. This peptide has one turn comprised of Ala-Pro-Asp-Arg, with the proline peptide bond in the cis configuration, and another turn that contains the disulfide and adjacent residues. If the disulfide is replaced by a thioether, and the central proline by an alpha-methyl proline, in an attempt to make the peptide more biologically stable, there is little adverse effect on affinity for ACA. The thioether bond/turn is fairly well defined with a Calpha to Calpha separation of 4.9 +/- 0.8 A. The alpha-methyl proline adopts the trans configuration, and this central Ala-(alpha-methyl-Pro)-Asp-Arg turn adopts a distorted type I turn conformation with a probable i to i+3 hydrogen bond. Modeling studies suggest that the proline peptide bond configuration switched from cis to trans in the presence of the alpha-methyl group on proline because of steric hindrance with the beta-carbon of the preceding residue. Overall, this peptidomimetic molecule is structurally very similar to the peptide with natural amino acids, with an rmsd difference of only 1.37 A, when comparing backbone atoms.