Spin Doctors

Magnetic resonance molecular imaging (MRMI), which uses the behavior of spinning protons under applied magnetic fields for visualizing specific factors in living tissue, appeals to both sides of the translational medicine divide. On the more basic side, precise detection of molecular targets promises detailed mechanistic information; in the clinic, MRMI holds potential for noninvasive diagnostics. For the technique to fulfill the hopes of M.D.s and Ph.D.s alike, straight-forward ways to make reliable probes is key. One tried-and-true strategy for generating a probe that can selectively detect a target molecule is phage display. Once a high-affinity binding sequence is identified, the corresponding peptide can be synthesized and coupled to iron oxide particles in order to create a magnetic reporter. However, it would be more convenient to use the selected phage directly, a concept that is explored in a recent paper from Segers et al,. The authors had previously selected phage that bind phosphatidylserine, an indicator of apoptosis. First, the phage were treated with ultrasmall particles of iron oxide that had been pretreated with epichlorhydrin. These paramagnetic particles could then react with amino groups present in the phage coat proteins, thus conjugating an average of 80 iron oxide particles to each phage virion. The authors tested the resulting “magnetophage” in a variety of in vitro experiments to determine binding avidity and selectivity. Although the apparent association constant was slightly lower when the magnetophage was compared to its standard counterpart, the MR-detectable version did show selectivity for phosphatidylserine as measured by competition assays with annexin V and as compared to interactions with phosphatidycholine. In vitro analyses included a magnetic ELISA-like assay that suggested that phosphatidyserine could be detected on apoptotic cells. Because of the potential relevance of MRMI to in vivo diagnostics, magnetophage were injected into mice, some of which had been treated to induce apoptosis of hepatic cells. The authors observed a significant difference in signal that correlated with the injection of phosphatidylserine-specific, pegylated magnetophage into mice with apoptotic livers. In this way, magnetophage may represent a convenient means to develop MRI contrast agents.

- Segers et al. 2007. From phage display to magnetophage, a new tool for magnetic resonance molecular imaging. Journal of Bioconjugate Chemistry [Epub ahead of print, May 24, 2007].

Tour de Force

Members of the Naval Research Laboratory have been actively developing innovative biosensors for optical detection of analyte-bound microbeads. Now, as advances in microfluidics and signal capture accelerate, Mulvaney et al,. report a biosensor capable of rapid, multiplexed assays that do not require amplification procedures or nanofabrication. Suitably enough for a research group based at a military institution, force is the basis of the assay. Molecular binding events occur on the picoNewton scale—the key is to tug just hard enough that nonspecific interactions are ruptured while label-analyte bonds remain intact. This selective disruption is achieved via laminar flow in microchannels and is called fluidic force discrimination (FFD). As conceived by Mulvaney et al,., the equipment consists of a microscope slide conjugated with capture oligonucleotides or antibodies. The slide forms the base of a flow cell into which the sample can be introduced for a 5–15-min incubation. This is followed by a sandwich-based detection scheme which culminates with addition of an oligo- or antibody-coupled 2.8-µm bead and FFD-enabling flow. Image capture via an optical microscope-mounted camera and spot recognition software delivers a microbead count, which is proportional to the analyte concentration. The authors also describe using a magnetoelectronic sensor designed for detecting the changes in electrical resistance that occur when captured paramagnetic microbeads are subjected to a magnetic field. Mulvaney et al,. demonstrate highly selective binding and an impressive dynamic range for detection of a 54-nucleotide oligo, ricin A chain, and staphylococcal enterotoxin B. A particularly appealing variant of their technique permits simultaneous nucleic acid and protein detection, which could add extra confidence to pathogen detection efforts. The team is currently working on converting the system to a field-deployable version that they envision being used for bioweapons detection, medical diagnostics, and environmental monitoring.

- Mulvaney et al. 2007. Rapid, femtomolar bioassays in complex matrices combining microfluidics and magnetoelectronics. Biosensors and Bioelectronics [Epub ahead of print, April 8, 2007].

Blue Clues

Just as in the macro world, complicated projects at the molecular level require a team of specialists. Discovering the exact composition of multiprotein complexes gives vital information about protein-protein networks, but the insights do not come easy. Crystallography and NMR can reveal the stoichiometry of proteins within a complex, but from material preparation to data analysis, determining atomic structure is a demanding process. Other strategies for detecting subunit ratios rely on metabolic labeling, FRET, or quantitative flow cytometry, but each of these methods has troublesome limitations. In a recent article, Swamy et al,. elected to try the much simpler technology of blue native-PAGE (BN-PAGE). In BN-PAGE, proteins are mixed with Coomassie blue dye, which binds nonspecifically to proteins and gives them an overall negative charge. However, unlike SDS, the dye does not disrupt protein-protein interactions and so the complex can be run intact on a native gradient gel. When antibodies against different members of the complex are added independently prior to electrophoresis, the resulting shift reveals how many antibodies are bound, and therefore how many molecules of that particular subunit are found in the complex. The assay was first tested with the B cell antigen receptor, which has a well-defined stoichiometry. As expected, a subunit that exists as only one copy produced a single shift, while subunits present at two copies produced two sequential shifts as more antibody was added. The authors then tested the strategy with the T cell antigen receptor, whose exact stoichiometry is debated. Although the banding pattern produced by some antibodies was inconclusive, by combining existing knowledge with unambiguous BN-PAGE findings, the authors felt comfortable reporting a complete stoichiometry for digitonin-solubilized T cell receptor. In other data, the authors show that unpurified protein complexes can be used in the assay, and that antibodies against HA and similar tags give good results against tagged proteins if submit-targeted antibodies are not available. With this technique, Coomassie blue offers a whole new means to understand protein behavior.

- Swamy et al. 2007. A native antibody-based mobility-shift technique (NAMOS-assay) to determine the stoichiometry of multiprotein complexes. Journal of Immunological Methods [Epub ahead of print, June 4, 2007].

No-Sweat FRET

One of the most spectacular ways to assay cellular events is through FRET-based biosensors. The basis of this technique is a protein or protein pair that undergoes a conformational change when the event to be detected occurs. Fluorescent proteins that can act as a FRET donor and acceptor are fused to the natural sensing protein such that the conformational change leads to excitation of the acceptor and fluorescence signal. The success of a FRET-based sensor depends on the dynamic range of the donor-acceptor pair, which itself follows from the conformational changes brought about when the biosensing protein is triggered. To improve the chances of creating a useful biosensor, Pham et al,. introduce FPMOD, the Fusion Protein Modeler. The program requires PDB structure files for the donor fluorophore (often CFP), the acceptor (usually YFP), and the biosensing protein (or protein pair) that serves as the backbone. The program begins by virtually joining these components with flexible linkers, and then models the effects of inserting or deleting linker residues on the fusion protein's conformational space. After removing sterically unfavorable possibilities, FPMOD calculates FRET efficiency changes upon conformational change for each structural variant. These in silico calculations should predict which fusion strategy is most likely to produce ligand-dependent FRET signal. As a test, Pham et al,. fed the program details of Calmodulin-based biosensors they had constructed for imaging of Ca²⁺ flux. The predicted FRET efficiencies were then tested in vitro, which revealed that the modeled behavior was consistent with the direction of FRET efficiency change, though not its absolute value. The authors then extended these successes by predicting optimal linker design for a Ca2+ sensor based on cadherin repeats. Future refinements of FPMOD might involve adding parameters to take into account Van der Waals interactions, solvation effects, and other properties that affect structural changes. Even so, the current version of FPMOD already demonstrates the potential of in silico-based rational design to minimize costly wet-lab optimization cycles.

- Pham et al. 2007. A computational tool for designing FRET protein biosensors by rigid-body sampling of their conformational space. Structure 15:515-523.