Search tips
Search criteria

Results 1-25 (245)

Clipboard (0)
Year of Publication
more »
1.  The GenK-catalyzed C-6′ Methylation in the Biosynthesis of Gentamicin: Isolation and Characterization of a Cobalamin-dependent Radical SAM Enzyme 
The existence of cobalamin (Cbl)-dependent enzymes that are members of the radical S-adenosyl-L-methionine (SAM) superfamily was previously predicted based on bioinformatic analysis. A number of these are Cbl-dependent methyltransferases but the details surrounding their reaction mechanisms have remained unclear. In this report we demonstrate the in vitro activity of GenK, a Cbl-dependent radical SAM enzyme that methylates an unactivated sp3 carbon during the biosynthesis of gentamicin, an aminoglycoside antibiotic. Experiments to investigate the stoichiometry of the GenK reaction revealed that one equivalent each of 5′-deoxyadenosine and S-adenosyl-homocysteine are produced for each methylation reaction catalyzed by GenK. Furthermore, isotope-labeling experiments demonstrate that the S-methyl group from SAM is transferred to Cbl and the aminoglycoside product during the course of the reaction. Based on these results, one mechanistic possibility for the GenK reaction can be ruled out and further questions regarding the mechanisms of Cbl-dependent radical SAM methyltransferases, in general, are discussed.
PMCID: PMC3796153  PMID: 23679096
2.  Diastereomeric Spirooxindoles as Highly Potent and Efficacious MDM2 Inhibitors 
Small-molecule inhibitors that block the MDM2-p53 protein-protein interaction (MDM2 inhibitors) are being intensely pursued as a new therapeutic strategy for cancer treatment. We previously published a series of spirooxindole-containing compounds as a new class of MDM2 small-molecule inhibitors. We report herein a reversible ring opening-cyclization reaction for some of these spirooxindoles, which affords four diastereomers from a single compound. Our biochemical binding data showed that the stereo-chemistry in this class of compounds has a major effect on their binding affinities to MDM2; with >100-fold difference between the most potent and the least potent stereoisomers. Our study has led to the identification of a set of highly potent MDM2 inhibitors with a stereochemistry that is different from that of our previously reported compounds. The most potent compound (MI-888) binds to MDM2 with a Ki value of 0.44 nM and achieves complete and long-lasting tumor regression in an animal model of human cancer.
PMCID: PMC3806051  PMID: 23641733
3.  Analysis of UDP-D-Apiose/UDP-D-Xylose Synthase-Catalyzed Conversion of UDP-D-Apiose Phosphonate to UDP-D-Xylose Phosphonate: Implications for a Retroaldol-Aldol Mechanism 
Journal of the American Chemical Society  2012;134(34):13946-13949.
UDP-D-apiose/UDP-D-xylose synthase (AXS) catalyzes the conversion of UDP-D-glucuronic acid to UDP-D-apiose and UDP-D-xylose. An acetyl-protected phosphonate analogue of UDP-D-apiose was synthesized and used in an in situ HPLC assay to demonstrate, for the first time, the ability of AXS to interconvert the two reaction products. Density functional theory calculations provided insight into the energetics of this process and the apparent inability of AXS to catalyze the conversion of UDP-D-xylose to UDP-D-apiose. The data suggest that this observation is unlikely to be due to an unfavorable equilibrium, but rather substrate inhibition by the most stable chair conformation of UDP-D-xylose. The detection of xylose cyclic phosphonate as the turnover product uncovers significant new detail about the AXS-catalyzed reaction and supports the proposed retroaldol-aldol mechanism of catalysis.
PMCID: PMC3454503  PMID: 22830643
UDP-D-apiose/UDP-D-xylose synthase; phosphonate substrate analogues; catalytic mechanism
4.  Inter-enzyme substrate diffusion for an enzyme cascade organized on spatially addressable DNA nanostructures 
Spatially addressable DNA nanostructures facilitate the self-assembly of heterogeneous elements with precisely controlled patterns. Here we organized discrete GOx/HRP enzyme pairs on specific DNA origami tiles with controlled inter-enzyme spacing and position. The distance between enzymes was systematically varied from 10 nm to 65 nm and the corresponding activities were evaluated. The study revealed two different distance dependent kinetic processes associated with the assembled enzyme pairs. Strongly enhanced activity was observed for those assemblies in which the enzymes were closely spaced, while the activity dropped dramatically for enzymes as little as 20 nm apart. Increasing the spacing further resulted in a much weaker distance dependence. Combined with diffusion modeling, the results suggest that Brownian diffusion of intermediates in solution governed the variations in activity for more distant enzyme pairs, while dimensionally-limited diffusion of intermediates across connected protein surfaces contributed to the enhancement in activity for closely spaced GOx/HRP assemblies. To further test the role of limited dimensional diffusion along protein surfaces, a noncatalytic protein bridge was inserted between GOx and HRP to connect their hydration shells. This resulted in substantially enhanced activity of the enzyme pair.
PMCID: PMC3319985  PMID: 22414276
5.  Engineering Polymeric Aptamers for Selective Cytotoxicity 
Journal of the American Chemical Society  2011;133(34):13380-13386.
Chemotherapy strategies thus far reported can result in both side effects and drug resistance. To address both of these issues at the cellular level, we report a molecular engineering strategy which employs polymeric aptamers to induce selective cytotoxicity inside target cells. The polymeric aptamers, composed of both multiple cell-based aptamers and a high ratio of dye-labeled short DNA, exploit the target recognition capability of the aptamer, enhanced cell internalization via multivalent effects, and cellular disruption by the polymeric conjugate. Importantly, the polymer backbone built into the conjugate is cytotoxic only inside cells. As a result, selective cytotoxicity is achieved equally in both normal cancer cells and drug-resistant cells. Control assays have confirmed the nontoxicity of the aptamer itself, but they have also shown that the physical properties of the polymer backbone contribute to target cell cytotoxicity. Therefore, our approach may shed new light on drug design and drug delivery.
PMCID: PMC3190636  PMID: 21702469
aptamers; polymerization; cytotoxicity
6.  Proteome Interrogation Using Nanoprobes to Identify Targets of a Cancer-killing Molecule 
We report a generic approach for identification of target proteins of therapeutic molecules using nanoprobes. Nanoprobes verify the integrity of nanoparticle-bound ligands in live cells and pull down target proteins from the cellular proteome, providing very important information on drug targets and mechanisms of action. As an example, target proteins for α-tubulin and HSP 90 were identified and validated.
PMCID: PMC3138543  PMID: 21495622
7.  Dynamic Presentation of Immobilized Ligands Regulated through Biomolecular Recognition 
Journal of the American Chemical Society  2010;132(39):13630-13632.
To mimic the dynamic regulation of signaling ligands immobilized on extracellular matrices or on the surfaces of neighboring cells for guidance of cell behavior and fate selection, we have harnessed biomolecular recognition in combination with polymer engineering to create dynamic surfaces on which the accessibility of immobilized ligands to cell surface receptors can be reversibly interconverted under physiological conditions. The cell-adhesive RGD peptide is chosen as a model ligand. RGD is fused to the C-terminus of a leucine zipper domain A and this fusion polypeptide is immobilized on surfaces through a residue at the N-terminus. The immobilized RGD can be converted from a cell-accessible to a cell-inaccessible state by addition of a conjugate of poly(ethylene) glycol (PEG) and another leucine zipper domain B (B-PEG). Heterodimerization between A and B allows co-immobilization of the PEG, which shields RGD from access by cells. The shielded RGD can be converted back to a cell-accessible state by addition of non-immobilized polypeptide A, which competes with the immobilized A for binding to B-PEG and removes B-PEG from the surface. This molecular design offers several advantages: the interconversion is reversible; the ligand remains immobilized during dynamic regulation so that cells are not exposed to the soluble form of the ligand that potentially has detrimental effects; the precision of the on/off states is assured by the molecular-level uniformity of the ligand and PEG co-immobilized through leucine zipper heterodimerization. The method can be readily adapted for dynamic regulation of other immobilized bioactive ligands of interest.
PMCID: PMC2992805  PMID: 20839864
9.  Molecular Behavior of DNA Origami in Higher-Order Self-Assembly 
Journal of the American Chemical Society  2010;132(38):13545-13552.
DNA-based self-assembly is a unique method for achieving higher-order molecular architectures made possible by the fact that DNA is a programmable information-coding polymer. In the past decade, two main types of DNA nanostructures have been developed: branch-shaped DNA tiles with small dimensions (commonly up to ~20 nm) and DNA origami tiles with larger dimensions (up to ~100 nm). Here we aimed to determine the important factors involved in the assembly of DNA origami superstructures. We constructed a new series of rectangular-shaped DNA origami tiles in which parallel DNA helices are arranged in a zigzag pattern when viewed along the DNA helical axis, a design conceived in order to relax an intrinsic global twist found in the original planar, rectangular origami tiles. Self-associating zigzag tiles were found to form linear arrays in both diagonal directions, while planar tiles showed significant growth in only one direction. Although the series of zigzag tiles were designed to promote two-dimensional array formation, one-dimensional linear arrays and tubular structures were observed instead. We discovered that the dimensional aspect ratio of the origami unit tiles and intertile connection design play important roles in determining the final products, as revealed by atomic force microscopy imaging. This study provides insight into the formation of higher-order structures from self-assembling DNA origami tiles, revealing their unique behavior in comparison with conventional DNA tiles having smaller dimensions.
PMCID: PMC3071357  PMID: 20825190
11.  Multilayer DNA Origami Packed on a Square Lattice 
Molecular self-assembly using DNA as a structural building block has proven to be an efficient route to the construction of nanoscale objects and arrays of increasing complexity. Using the remarkable “scaffolded DNA origami” strategy, Rothemund demonstrated that a long single-stranded DNA from a viral genome (M13) can be folded into a variety of custom two-dimensional (2D) shapes using hundreds of short synthetic DNA molecules as staple strands. More recently, we generalized a strategy to build custom-shaped, three-dimensional (3D) objects formed as pleated layers of helices constrained to a honeycomb lattice, with precisely controlled dimensions ranging from 10 to 100 nm. Here we describe a more compact design for 3D origami, with layers of helices packed on a square lattice, that can be folded successfully into structures of designed dimensions in a one-step annealing process, despite the increased density of DNA helices. A square lattice provides a more natural framework for designing rectangular structures, the option for a more densely packed architecture, and the ability to create surfaces that are more flat than is possible with the honeycomb lattice. Thus enabling the design and construction of custom 3D shapes from helices packed on a square lattice provides a general foundational advance for increasing the versatility and scope of DNA nanotechnology.
PMCID: PMC2821935  PMID: 19807088
14.  New Reactions and Products Resulting from Alternative Interactions between the P450 Enzyme and Redox Partners 
Cytochrome P450 enzymes are capable of catalyzing a great variety of synthetically useful reactions such as selective C–H functionalization. Surrogate redox partners are widely used for reconstitution of P450 activity based on the assumption that the choice of these auxiliary proteins or their mode of action does not affect the type and selectivity of reactions catalyzed by P450s. Herein, we present an exceptional example to challenge this postulate. MycG, a multifunctional biosynthetic P450 monooxygenase responsible for hydroxylation and epoxidation of 16-membered ring macrolide mycinamicins, is shown to catalyze the unnatural N-demethylation(s) of a range of mycinamicin substrates when partnered with the free Rhodococcus reductase domain RhFRED or the engineered Rhodococcus-spinach hybrid reductase RhFRED-Fdx. By contrast, MycG fused with the RhFRED or RhFRED-Fdx reductase domain mediates only physiological oxidations. This finding highlights the larger potential role of variant redox partner protein–protein interactions in modulating the catalytic activity of P450 enzymes.
PMCID: PMC3985502  PMID: 24521145
15.  Mechanistic Consequences of Chiral Radical Clock Probes: Analysis of the Mononuclear Non-Heme Iron Enzyme HppE with 2-Hydroxy-3-methylenecyclopropyl Radical Clock Substrates 
(S)-2-Hydroxypropylphosphonic acid [(S)-HPP] epoxidase (HppE) is a mononuclear iron enzyme that catalyzes the last step in the biosynthesis of the antibiotic fosfomycin. HppE also processes the (R)-enantiomer of HPP but converts it to 2-oxo-propylphosphonic acid. In this study, all four stereoisomers of 3-methylenecyclopropyl-containing substrate analogues, (2R, 3R)-8, (2R, 3S)-8, (2S, 3R)-8, and (2S, 3S)-8, were synthesized and used as radical probes to investigate the mechanism of the HppE-catalyzed reaction. Upon treatment with HppE, (2S, 3R)-8 and (2S, 3S)-8 were converted via a C1 radical intermediate to the corresponding epoxide products, as anticipated. In contrast, incubation of HppE with (2R, 3R)-8 led to enzyme inactivation, and incubation of HppE with (2R, 3S)-8 yielded the 2-keto product. The former finding is consistent with the formation of a C2 radical intermediate, where the inactivation is likely triggered by radical-induced ring cleavage of the methylenecyclopropyl group. Reaction with (2R, 3S)-8 is predicted to also proceed via a C2 radical intermediate, but no enzyme inactivation and no ring-opened product were detected. These results strongly suggest that an internal electron transfer to the iron center subsequent to C–H homolysis competes with ring-opening in the processing of the C2 radical intermediate. The different outcomes of the reactions with (2R, 3R)-8 and (2R, 3S)-8 demonstrate the need to carefully consider the chirality of substituted cyclopropyl groups as radical reporting groups in studies of enzymatic mechanisms.
PMCID: PMC4004275  PMID: 24512048
16.  Single siRNA Nanocapsules for Enhanced RNAi Delivery 
Journal of the American Chemical Society  2012;134(33):13542-13545.
Synthetic siRNA has been considered as a highly promising therapeutic agent for human diseases. However, clinical use of siRNA has been hampered by instability in the body and inability to deliver sufficient RNA interference compounds to the tissues or cells. To address this challenge, we present here a single siRNA nanocapsule delivery technology, which is achieved by encapsulating a single siRNA molecule within a degradable polymer nanocapsule with a diameter around 20 nm and positive surface charge. As proof-of-concept, since CCR5 is considered a major silencing target of HIV therapy, CCR5–siRNA nanocapsules were delivered into 293T cells and successfully downregulated the CCR5 RNA fused with mCherry reporter RNA. In the absence of human serum, nanocapsules and lipofectamine silenced expression of CCR5–mCherry expression to 8% and 15%, respectively. Such nanocapsules maintain the integrity of siRNA inside even after incubation with ribonuclease and serum for 1 h; under the same conditions, siRNA is degraded in the native form or when formulated with lipofectamine. In the presence of serum, CCR5–siRNA nanocapsules knocked down CCR5–mCherry expression to less than 15% while siRNAs delivered through lipofectamine slightly knocked down the expression to 55%. In summary, this work provides a novel platform for siRNA delivery that can be developed for therapeutic purposes.
PMCID: PMC4318836  PMID: 22866878
18.  The Active Site Sulfenic Acid Ligand in Nitrile Hydratases Can Function as a Nucleophile 
Nitrile hydratase (NHase) catalyzes the hydration of nitriles to their corresponding commercially valuable amides at ambient temperatures and physiological pH. Several reaction mechanisms have been proposed for NHase enzymes however, the source of the nucleophile remains a mystery. Boronic acids have been shown to be potent inhibitors of numerous hydrolytic enzymes due to the open shell of boron, which allows it to expand from a trigonal planar (sp2) form to a tetrahedral form (sp3). Therefore, we examined the inhibition of the Co-type NHase from Pseudonocardia thermophila JCM 3095 (PtNHase) by boronic acids via kinetics and X-ray crystallography. Both 1-butaneboronic acid (BuBA) and phenylboronic acid (PBA) function as potent competitive inhibitors of PtNHase. X-ray crystal structures for BuBA and PBA complexed to PtNHase were solved and refined at 1.5, 1.6 and 1.2 Å resolution. The resulting PtNHase-boronic acid complexes represent a “snapshot” of reaction intermediates and implicate the cysteine-sulfenic acid ligand as the catalytic nucleophile, a heretofore unknown role for the αCys113-OH sulfenic acid ligand. Based on these data, a new mechanism of action for the hydration of nitriles by NHase is presented.
PMCID: PMC3968781  PMID: 24383915
19.  Structure and Inhibition of Tuberculosinol Synthase and Decaprenyl Diphosphate Synthase from Mycobacterium tuberculosis 
We have obtained the structure of the bacterial diterpene synthase, tuberculosinol/iso-tuberculosinol synthase (Rv3378c) from Mycobacterium tuberculosis, a target for anti-infective therapies that block virulence factor formation. This phosphatase adopts the same fold as found in the Z- or cis-prenyltransferases. We also obtained structures containing the tuberculosinyl diphosphate substrate together with one bisphosphonate inhibitor-bound structure. These structures together with the results of site-directed mutagenesis suggest an unusual mechanism of action involving two Tyr residues. Given the similarity in local and global structure between Rv3378c and the M. tuberculosis cis-decaprenyl diphosphate synthase (DPPS; Rv2361c), the possibility exists for the development of inhibitors that target not only virulence but also cell wall biosynthesis, based in part on the structures reported here.
PMCID: PMC3986019  PMID: 24475925
20.  In vitro Characterization of LmbK and LmbO: Identification of GDP-D-erythro-α-D-gluco-Octose as a Key Intermediate in Lincomycin A Biosynthesis 
Lincomycin A is a clinically useful antibiotic isolated from Streptomyces lincolnensis. It contains an unusual methylmercapto-substituted octose, methylthiolincosamide (MTL). While it has been demonstrated that the C8 backbone of MTL moiety is derived from D-fructose 6-phosphate and D-ribose 5-phosphate via a transaldol reaction catalyzed by LmbR, the subsequent enzymatic transformations leading to the MTL moiety remain elusive. Here, we report the identification of GDP-D-erythro-α-D-gluco-octose (GDP-D-α-D-octose) as a key intermediate in the MTL biosynthetic pathway. Our data show that the octose 1,8-bisphosphate intermediate is first converted to octose 1-phosphate by a phosphatase, LmbK. The subsequent conversion of the octose 1-phosphate to GDP-D-α-D-octose is catalyzed by the octose 1-phosphate guanylyltransferase, LmbO. These results provide significant insight into the lincomycin biosynthetic pathway, because the activated octose likely serves as the acceptor for the installation of the C1 sulfur appendage of MTL.
PMCID: PMC3932030  PMID: 24380627
21.  Ligand Concentration Regulates the Pathways of Coupled Protein Folding and Binding 
Coupled ligand binding and conformational change plays a central role in biological regulation. Ligands often regulate protein function by modulating conformational dynamics, yet the order in which binding and conformational change occurs are often hotly debated. Here we show that the “conformational selection versus induced fit” on which this debate is based is a false dichotomy because the mechanism depends on ligand concentration. Using the binding of pyrophosphate (PPi) to B. subtilis RNase P protein as a model, we show that coupled reactions are best understood as a change in flux between competing pathways with distinct orders of binding and conformational change. The degree of partitioning through each pathway depends strongly on PPi concentration, with ligand binding redistributing the conformational ensemble toward the folded state by both increasing folding rates and decreasing unfolding rates. These results indicate that ligand binding induces marked and varied changes in protein conformational dynamics, and that the order of binding and conformational change is ligand concentration dependent.
PMCID: PMC3977005  PMID: 24364358
22.  Observation of Complete Pressure-Jump Protein Refolding in Molecular Dynamics Simulation and Experiment 
Density is an easily adjusted variable in molecular dynamics (MD) simulations. Thus, pressure-jump (P-jump)-induced protein refolding, if it could be made fast enough, would be ideally suited for comparison with MD. Although pressure denaturation perturbs secondary structure less than temperature denaturation, protein refolding after a fast P-jump is not necessarily faster than that after a temperature jump. Recent P-jump refolding experiments on the helix bundle λ-repressor have shown evidence of a <3 μs burst phase, but also of a ∼1.5 ms “slow” phase of refolding, attributed to non-native helical structure frustrating microsecond refolding. Here we show that a λ-repressor mutant is nonetheless capable of refolding in a single explicit solvent MD trajectory in about 19 μs, indicating that the burst phase observed in experiments on the same mutant could produce native protein. The simulation reveals that after about 18.5 μs of conformational sampling, the productive structural rearrangement to the native state does not occur in a single swift step but is spread out over a brief series of helix and loop rearrangements that take about 0.9 μs. Our results support the molecular time scale inferred for λ-repressor from near-downhill folding experiments, where transition-state population can be seen experimentally, and also agrees with the transition-state transit time observed in slower folding proteins by single-molecule spectroscopy.
PMCID: PMC3985862  PMID: 24437525
23.  Quantum Dots Encapsulated within Phospholipid Membranes: Phase-Dependent Structure, Photostability, and Site-Selective Functionalization 
Lipid vesicle encapsulation is an efficient approach to transfer quantum dots (QDs) into aqueous solutions, which is important for renewable energy applications and biological imaging. However, little is known about the molecular organization at the interface between a QD and lipid membrane. To address this issue, we investigated the properties of 3.0 nm CdSe QDs encapsulated within phospholipid membranes displaying a range of phase transition temperatures (Tm). Theoretical and experimental results indicate that the QD locally alters membrane structure, and in turn, the physical state (phase) of the membrane controls the optical and chemical properties of the QDs. Using photoluminescence, ICP-MS, optical microscopy, and ligand exchange studies, we found that the Tm of the membrane controls optical and chemical properties of lipid vesicle-embedded QDs. Importantly, QDs encapsulated within gel-phase membranes were ultrastable, providing the most photostable non-core/shell QDs in aqueous solution reported to date. Atomistic molecular dynamics simulations support these observations and indicate that membranes are locally disordered displaying greater disordered organization near the particle–solution interface. Using this asymmetry in membrane organization near the particle, we identify a new approach for site-selective modification of QDs by specifically functionalizing the QD surface facing the outer lipid leaflet to generate gold nanoparticle–QD assemblies programmed by Watson–Crick base-pairing.
PMCID: PMC3985776  PMID: 24417287
24.  The Active Site Sulfenic Acid Ligand in Nitrile Hydratases Can Function as a Nucleophile 
Nitrile hydratase (NHase) catalyzes the hydration of nitriles to their corresponding commercially valuable amides at ambient temperatures and physiological pH. Several reaction mechanisms have been proposed for NHase enzymes; however, the source of the nucleophile remains a mystery. Boronic acids have been shown to be potent inhibitors of numerous hydrolytic enzymes due to the open shell of boron, which allows it to expand from a trigonal planar (sp2) form to a tetrahedral form (sp3). Therefore, we examined the inhibition of the Co-type NHase from Pseudonocardia thermophila JCM 3095 (PtNHase) by boronic acids via kinetics and X-ray crystallography. Both 1-butaneboronic acid (BuBA) and phenylboronic acid (PBA) function as potent competitive inhibitors of PtNHase. X-ray crystal structures for BuBA and PBA complexed to PtNHase were solved and refined at 1.5, 1.6, and 1.2 Å resolution. The resulting PtNHase–boronic acid complexes represent a “snapshot” of reaction intermediates and implicate the cysteine-sulfenic acid ligand as the catalytic nucleophile, a heretofore unknown role for the αCys113–OH sulfenic acid ligand. Based on these data, a new mechanism of action for the hydration of nitriles by NHase is presented.
PMCID: PMC3968781  PMID: 24383915
25.  Superhydrophobic Photosensitizers. Mechanistic Studies of 1O2 Generation in the Plastron and Solid/Liquid Droplet Interface 
Journal of the American Chemical Society  2013;135(50):18990-18998.
We describe here a physical-organic study of the first triphasic superhydrophobic sensitizer for photooxidations in water droplets. Control of synthetic parameters enables the mechanistic study of “borderline” two- and three-phase superhydrophobic sensitizer surfaces where 1O2 is generated in compartments that are wetted, partially wetted, or remain dry in the plastron (i.e., air layer beneath the droplet). The superhydrophobic surface is synthesized by partially embedding silicon phthalocyanine (Pc) sensitizing particles to specific locations on polydimethylsiloxane (PDMS) posts printed in a square array (1 mm tall posts on 0.5 mm pitch). In the presence of red light and oxygen, singlet oxygen is formed on the superhydrophobic surface and reacts with 9,10-anthracene dipropionate dianion (1) within a freestanding water droplet to produce an endoperoxide in 54–72% yields. Control of the 1O2 chemistry was achieved by the synthesis of superhydrophobic surfaces enriched with Pc particles either at the PDMS end-tips or at PDMS post bases. Much of the 1O2 that reacts with anthracene 1 in the droplets was generated by the sensitizer “wetted” at the Pc particle/water droplet interface and gave the highest endoperoxide yields. About 20% of the 1O2 can be introduced into the droplet from the plastron. The results indicate that the superhydrophobic sensitizer surface offers a unique system to study 1O2 transfer routes where a balance of gas and liquid contributions of 1O2 is tunable within the same superhydrophobic surface.
PMCID: PMC3930608  PMID: 24295210

Results 1-25 (245)