SHAPE Shifter

Magical explanations thrive where concrete knowledge fails, and so RNA crystallographers can be forgiven if they sometimes suspect that the phase of the moon affects crystal formation and quality. As of yet, rational engineering of RNA molecules to improve diffraction is not possible. Some tricks of the trade are known—adding/deleting nucleotides at the ends of the molecule and stabilizing internal structural elements—but the overall process remains one of trial and error. As a result, a method that allowed crystallographers to predict which nucleotides undergo conformational changes during crystallization would greatly assist efforts to rationally design crystallization-competent RNAs. To that end, Vicens et al., describe the application of a structure-probing technique called selective 2′-hydroxyl acylation analyzed by primer extension (SHAPE) to the mapping of the reactivity and flexibility of nucleotides in RNAs designated for crystallization. As the name implies, the method probes the reactivity of a nucleotide's 2′ hydroxyl, allowing the user to distinguish between a nucleotide that is inert because it is already involved in base pairing versus one that is accessible to the reactant. By comparing SHAPE results obtained in the lattice-bound RNA with the findings obtained when the molecule is in native solution or crystallization buffer, researchers can see which nucleotides shift as they form intra- and intermolecular contacts. These residues are most likely to be critical to crystal quality and therefore represent the best targets for site-directed mutagenesis or in vitro selection strategies. The authors also suggest that SHAPE could be used even when no crystal is available, since useful information could likely be deduced if RNA behavior were monitored across a variety of crystallization conditions. Although SHAPE may not yet be the magic ticket to a high-quality RNA crystal, its detailed mapping of structural behavior will undoubtedly help guide the way.

- Vicens et al. 2007. Local RNA structural changes induced by crystallization are revealed by SHAPE. RNA 13:536-548.

SherLoc is on the Trail

Like detectives in a miniature whodunit, cell biologists would love to be able to figure out exactly what each protein in the cell is doing and why. They know that one of the best sources of clues is to surveil the main suspects and try to infer, from their location, what precisely they're up to. Although there are a variety of ways to determine that information experimentally, these techniques are time-consuming and may not yield meaningful data for all proteins. In this context, the appeal of computational methods makes perfect sense. Most subcellular localization packages have relied on sequencebased rules. Now, Shatkay et al., introduce SherLoc, a system that uses sequence information and text mining to predict protein localization. This package is based on the premise that certain “distinguishing terms” correlate with particular subcellular localization patterns. For example, a protein name that often appears adjacent to the text fragment “histon” has an increased probability of referring to a nuclear protein; similarly, “cytochrom” might indicate a mitochondrial protein, while “chaperon” could imply a protein that is associated with the endoplasmic reticulum. In all, 11 possible localizations and about 550 distinguishing terms are tested for each query protein, and the highest probability match is combined with four sequence-based algorithms to return the final predicted localization. Shatkay et al., found that SherLoc's combination of sequence- and text-based prediction performed better than any of the components alone, and that the accuracy of localization was 70% or more. Extracellular proteins appeared to be easiest to place, while accurately distinguishing nuclear from cytoplasmic factors proved the most troublesome. For researchers interested in a hint about the favorite haunts of their proteins of interest, the authors also report SherLoc's verdict on 19,000 eukaryotic proteins with uncertain or unknown subcellular localizations; the results are available at www-bs.informatik.uni-tuebingen.de/Services/SherLoc, where the program is also available.

- Shatkay et al. 2007. SherLoc: high-accuracy prediction of protein subcellular localization by integrating text and protein sequence data. Bioinformatics [Epub ahead of print, April 9, 2007].

Diamonds Are For…

People have often turned to diamonds when they find their other half, so perhaps it's no surprise that diamond surfaces are playing a role in one of the most exquisitely perfect pairings known to nature: an antibody and its antigen. The famously robust gemstones are, not surprisingly, very chemically stable, which makes them an appealing platform for electrical impedance spectroscopy (EIS) detection of biomolecular binding events. EIS-based biosensing involves a biofunctionalized semiconducting surface that is exposed to an analyte in solution. If the target binds with the covalently immobilized probe, the charge of the newly proximate molecule perturbs conductivity sufficiently to induce an EIS signal. The result is label-free detection of binding events. Previous studies have shown the suitability of functionalized silicon and diamond surfaces for detecting nucleic acid binding events. In an extension of this promising work, Yang et al., describe the use of these surfaces in EIS monitoring of antigen-antibody interactions. In this proof-of-principle analysis, the authors used a photochemical process to functionalize diamond films or silicon surfaces with a layer of primary amines, to which IgG was covalently attached in a subsequent step. The analyte in this case was fluorescently labeled anti-human IgG, which was injected into an electrochemical fluid cell. The authors provide a detailed analysis of the impedance measurements and the electrical properties of both the diamond and silicon surfaces. In the data shown, the sensitivity of EIS was marginally less than that achievable with fluorescent detection. However, the authors point out that fluorescence measurements require washing of the reaction chamber, while EIS detection can occur in real time. Although the diamond-based films present some challenges compared with the more widely employed silicon surfaces, the authors emphasize that the results confirm the potential of this incredibly robust surface for real-time biosensing applications.

- Yang et al. 2007. Direct electrical detection of antigen-antibody binding on diamond and silicon substrates using electrical impedance spectroscopy. The Analyst 132:296-306.

PET Project

Fluorescently tagging a protein to visualize it in the living cell is a routine technique, a fact that must be a source of jealousy amongst researchers who find it much more difficult to spy on RNA. For RNA localization strategies, the closest analogue to GFP binding is to tag the transcript with recognition sites for the RNA binding protein of the MS2 bacteriophage. Co-expression of a GFP-MS2 binding protein fusion construct causes recruitment of the fluorescent moiety to the tagged transcript. This approach, though elegant, offers a signal-to-noise ratio that is not high enough for dynamic monitoring without extensive optimization. A pair of University of Pittsburgh researchers conjectured that an approach based on chemical biology might address this limitation, and in a paper appearing in JACS, they describe a PET-based RNA sensor. Photoinduced Electron Transfer (PET) sensors rely on reversible conformational changes that sequester a quencher from a fluorophore to activate the fluorescence signal. The authors hypothesized that they could trigger this conformational change by using in vitro RNA selection to identify an aptamer that binds specifically to the quencher. Straightforward cloning could be used to introduce the binding motif to a transcript of interest; as the RNA is produced in cells containing the small-molecule PET sensor, its quencher-binding portion would light up and signal the location of the chimeric RNA. For preliminary evidence that this appealing strategy has potential, the authors describe the synthesis of a 2′, 7′-dichloro-fluorescein fluorophore conjugated to dual aniline-based quenchers and its use as a target for in vitro selection. Using a couple of different RNA libraries, they amassed a collection of aptamers that activated fluorescence of the PET sensor by about 4-fold. While this level of activation remains far from the level that would be required for practical cell-based imaging, the results do confirm the validity of the approach and provide hope that this effort will yield a vibrant new tool for tracking RNA in living cells.

- Sparano and Koide. 2007. Fluorescent sensors for specific RNA: a general paradigm using chemistry and combinatorial biology. Journal of the American Chemical Society 129:4785-4794.