Scott J. Miller

Nymble: a High-Performance Learning Name-finder

This paper presents a statistical, learned approach to finding names and other nonrecursive entities in text (as per the MUC-6 definition of the NE task), using a variant of the standard hidden Markov model. We present our justification for the problem and our approach, a detailed discussion of the model itself and finally the successful results of this new approach.

Dynamic Kinetic Resolution of Biaryl Atropisomers via Peptide-Catalyzed Asymmetric Bromination

Selectively Spun Biaryl compounds, in which two phenyl rings are linked by a single bond, exhibit an interesting sort of chirality, termed atropisomerism. If bulky substituents block the mutual rotation of the rings about the linking bond, then two isomers can be isolated that differ only in the direction one ring has swiveled away from the plane of the other. This feature is useful in ligand design for asymmetric catalysis and also appears in a number of polycyclic natural products. However, selective synthesis of a single isomer is difficult. Gustafson et al. (p. 1251) now show that a simple tripeptide derivative acts as an efficient catalyst for this purpose, trapping a freely rotating precursor in one orientation through selective bromination; the large bromine substituents then inhibit further swiveling. A simple catalyst converts antibiotic-like molecules into near-single chiral form. Despite the widespread use of axially chiral, or atropisomeric, biaryl ligands in modern synthesis and the occurrence of numerous natural products exhibiting axial chirality, few catalytic methods have emerged for the direct asymmetric preparation of this compound class. Here, we present a tripeptide-derived small-molecule catalyst for the dynamic kinetic resolution of racemic biaryl substrates. The reaction proceeds via an atropisomer-selective electrophilic aromatic substitution reaction using simple bromination reagents. The result is an enantioselective synthesis that delivers chiral nonracemic biaryl compounds with excellent optical purity and good isolated chemical yields (in most cases a >95:5 enantiomer ratio and isolated yields of 65 to 87%). A mechanistic model is advanced that accounts for the basis of selectivity observed.

Iridium-Catalyzed Hydrogenation of N-Heterocyclic Compounds under Mild Conditions by an Outer-Sphere Pathway

A new homogeneous iridium catalyst gives hydrogenation of quinolines under unprecedentedly mild conditions-as low as 1 atm of H(2) and 25 °C. We report air- and moisture-stable iridium(I) NHC catalyst precursors that are active for reduction of a wide variety of quinolines having functionalities at the 2-, 6-, and 8- positions. A combined experimental and theoretical study has elucidated the mechanism of this reaction. DFT studies on a model Ir complex show that a conventional inner-sphere mechanism is disfavored relative to an unusual stepwise outer-sphere mechanism involving sequential proton and hydride transfer. All intermediates in this proposed mechanism have been isolated or spectroscopically characterized, including two new iridium(III) hydrides and a notable cationic iridium(III) dihydrogen dihydride complex. DFT calculations on full systems establish the coordination geometry of these iridium hydrides, while stoichiometric and catalytic experiments with the isolated complexes provide evidence for the mechanistic proposal. The proposed mechanism explains why the catalytic reaction is slower for unhindered substrates and why small changes in the ligand set drastically alter catalyst activity.

A Fully Statistical Approach to Natural Language Interfaces

We present a natural language interface system which is based entirely on trained statistical models. The system consists of three stages of processing: parsing, semantic interpretation, and discourse. Each of these stages is modeled as a statistical process. The models are fully integrated, resulting in an end-to-end system that maps input utterances into meaning representation frames.

Functional Analysis of an Aspartate‐Based Epoxidation Catalyst with Amide‐to‐Alkene Peptidomimetic Catalyst Analogues

The biosynthesis of natural products that contain epoxides represents a powerful stimulus for the study of “epoxidase” enzymes.[i] Likewise, these processes have inspired a generation of science focused on small molecule catalysts that mediate selective epoxidations through a variety of mechanisms.[ii] With respect to the naturally occurring epoxidases, the mechanistic basis of O-atom transfer is often associated with the chemistry of either flavinoid cofactors, P450 enzymes containing a heme group, or chloroperoxidases that lead to stepwise ring formation.[iii] In thinking about the known biosynthetic apparatus for epoxide formation, we became curious about an alternative mode for O-atom transfer – one based on functional groups available in proteins, but perhaps not well-documented in the biosynthesis of epoxides. In particular, we speculated and recently showed that aspartic-acid-containing peptides (e.g., 1; Figure 1a) might shuttle between the side-chain carboxylic acid and the corresponding peracid (e.g., 2) creating a catalytic cycle competent for asymmetric epoxidation with turnover of the aspartate-derived catalyst. Such an approach is orthogonal to the Julia-Colonna epoxidation, a complementary peptide-based epoxidation based on a nucleophilic mechanism.[iv] Indeed, as shown in Figure 1b, this new electrophilic epoxidation catalytic cycle mediates the asymmetric epoxidation of substrates like 3 to give products like 4 with up to 92% ee.[v]

Pyridylalanine (Pal)-Peptide Catalyzed Enantioselective Allenoate Additions to N-Acyl Imines

An amine-catalyzed reaction has been discovered that couples alpha-allenic esters with N-acyl imines in good to excellent yields (up to 88%). Extension of this methodology from the study of achiral pyridine-based catalysis to chiral peptide-based scaffolds is presented. The approach culminated in the identification of a tetrameric peptide sequence containing an embedded pyridylalanine (Pal) residue as an efficient asymmetric catalyst for enantioselective coupling reactions. The unique allenic products are obtained with enantiomer ratios of up to approximately 95:5 (up to >98:2 following recrystallization).

Incorporation of Peptide Isosteres into Enantioselective Peptide-Based Catalysts as Mechanistic Probes

Olefinic analogues of effective peptide-based catalysts for the kinetic resolution of functionalized racemic secondary alcohols have been synthesized. The isosteric replacement of the peptide amide bond in 1 with an (E)-alkene to form 2 has enabled the evaluation of the kinetic role of particular amides within these catalysts.

Enantioselective sulfonylation reactions mediated by a tetrapeptide catalyst

While Nature excels at performing selective modifications of complex polyfunctional molecules through the use of tailoring enzymes, synthetic chemistry has lagged behind in this regard. In prior work, we have applied a biomimetic approach to this problem, developing small peptides to achieve various group transfer reactions on polyol substrates with high enantio- or regioselectivity. The utility of sulfonates as synthetic building blocks and the scarcity of direct, selective methods for their preparation prompted our investigation into this area. In this article we report the development of a π-methyl histidine-based tetrameric peptide that effects the desymmetrization of meso-1,3-diols through enantioselective (mono)sulfonylation. The catalyst exhibits structural similarities to another catalyst found to be effective in orthogonal group transfers, but results in modification of the enantiotopic alcohol. The practical and mechanistic implications of this discovery may extend beyond synthetic considerations and provide analogies to the diverse roles of histidine in enzyme active sites.

n → π* Interaction and n)(π Pauli Repulsion Are Antagonistic for Protein Stability

In many common protein secondary structures, such as alpha-, 3(10), and polyproline II helices, an n --> pi* interaction places the adjacent backbone amide carbonyl groups in close proximity to each other. This interaction, which is reminiscent of the Burgi-Dunitz trajectory, involves delocalization of the lone pairs (n) of the oxygen (O(i-1)) of a peptide bond over the antibonding orbital (pi*) of C(i)=O(i) of the subsequent peptide bond. Such a proximal arrangement of the amide carbonyl groups should be opposed by the Pauli repulsion between the lone pairs (n) of O(i-1) and the bonding orbital (pi) of C(i)=O(i). We explored the conformational effects of this Pauli repulsion by employing common peptidomimetics, wherein the n --> pi* interaction is attenuated while the Pauli repulsion is retained. Our results indicate that this Pauli repulsion prevents the attainment of such proximal arrangement of the carbonyl groups in the absence of the n --> pi* interaction. This finding indicates that the poor mimicry of the amide bond by many peptidomimetics stems from their inability to partake in the n --> pi* interaction and emphasizes the quantum-mechanical nature of the interaction between adjacent amide carbonyl groups in proteins.

Determination of Noncovalent Docking by Infrared Spectroscopy of Cold Gas-Phase Complexes

Ties That Bind Almost by definition, effective catalysts bind their substrates for a very short time—releasing them quickly after helping them react and then moving on to bind new, as yet unreacted, substrates. This property engenders an efficient cycle, but it hinders study of the binding motif. Garand et al. (p. 694, published online 19 January; see the Perspective by Zwier) devised a technique to extract bound complexes from solution and freeze their conformations in cold, gas-phase clusters. Probing these clusters by vibrational spectroscopy in conjunction with theoretical calculations then allowed the sites of hydrogen bonding that hold the complexes together to be pinpointed. Conformationally freezing a weakly bound complex in the gas phase sheds light on its likely binding motifs in solution. Multidentate, noncovalent interactions between small molecules and biopolymer fragments are central to processes ranging from drug action to selective catalysis. We present a versatile and sensitive spectroscopic probe of functional groups engaged in hydrogen bonding in such contexts. This involves measurement of the frequency changes in specific covalent bonds upon complex formation, information drawn from otherwise transient complexes that have been extracted from solution and conformationally frozen near 10 kelvin in gas-phase clusters. Resonances closely associated with individual oscillators are easily identified through site-specific isotopic labeling, as demonstrated by application of the method to an archetypal system involving a synthetic tripeptide known to bind biaryl substrates through tailored hydrogen bonding to catalyze their asymmetric bromination. With such data, calculations readily converge on the plausible operative structures in otherwise computationally prohibitive, high-dimensionality landscapes.