Tag—you're it

It has been known for decades that one X chromosome is randomly inactivated in each cell in XX individuals, but the mechanism ensuring that only one chromosome is silenced has remained unclear. Na Xu and colleagues now provide the first evidence that direct interchromosomal pairing is at least part of the explanation for the mutually exclusive silencing of a single X chromosome in dosage-compensated cells (Science, published online 19 January 2006; 10.1126/science.1122984). The authors examined chromosome movement in differentiating mouse embryonic stem cells, which are known to recapitulate X-chromosome inactivation (XCI) in culture. Using fluorescence in situ hybridization, they showed that X-X interchromosomal distances narrow significantly during the period when XCI is known to occur. This 'proximity pairing' is restricted to the X-inactivation center, the cis-acting region that initiates chromosome-wide silencing. To show that this proximity reflects actual physical pairing, Xu et al. carried out chromosome conformation capture (3C), which detects intermolecular cross-links. Indeed, the 3C method consistently reveals interactions at the Tsix and Xite loci specifically, two elements that are also required in cis for XCI. Insertion of these elements onto autosomes promotes de novo X-autosome interactions, showing that either is sufficient to promote pairing. AP

Frizzled on the move

Establishment of planar cell polarity in the fly wing requires the asymmetric distribution of proteins at the proximal or distal edge of developing wing cells, but the cellular mechanisms driving these asymmetric localization events remain poorly defined. New work by Tadashi Uemura and colleagues (Dev. Cell 10, 209–222; 2006) has now identified a polarized array of microtubules located beneath the apical cell surface that appears to guide the directional transport of the transmembrane protein Frizzled, a core PCP component, to the distal edge of pupal wing cells. The authors used a green fluorescent protein tag to visualize Frizzled trafficking in live cells and found that particles carrying the tagged protein were enriched at the level of adherens junctions and moved preferentially toward distal cell boundaries. At the same position, they identified a network of microtubules preferentially oriented along the cell's proximal-distal axis and showed that the growing ends of the microtubules were enriched distally. The findings suggest a model whereby Frizzled-positive vesicles are transported vectorially along a distally oriented microtubule network by the activity of a plus end–directed motor protein, leading to the preferential accumulation of Frizzled at the distal cell membrane. KV

Comparative protein networks

The prediction of functional orthologs across species has traditionally been based on sequence or expression comparisons. Presenting an alternative approach, Trey Ideker and colleagues consider the use of network analyses (Genome Res. published online 27 January 2006; 10.1101/gr.4526003). Whereas other studies have compared protein interaction networks in order to identify architecture, regions or specific components that are common between species, the current study compares networks to predict functional orthologs. This is based on the tested hypothesis that proteins with conserved function are likely to retain many of the same interacting partners. At the core of their approach is the previously developed Inparanoid algorithm, which assigns sequence-based homology clusters, applied to the protein-protein interaction networks of Drosophila melanogaster and Saccharomyces cerevisiae. The authors then examined a network of interacting proteins, defining a 'conservation index' that reflects the proportion of protein interactions conserved between species for a particular protein pair, and developed a probabilistic method based on this index to estimate the likelihood of functional orthology. These methods were tested on 121 cases in which sequence comparisons alone left unclear predictions. In approximately half of these cases, the authors' network-based approach suggested that a different protein pair was the most likely functional ortholog, highlighting important differences from traditional comparative sequence approaches. OB

Specifying midbrain dopaminergic fate

The ability to instruct embryonic stem (ES) cells to differentiate as midbrain dopaminergic neurons holds promise as a potential cell replacement therapy for Parkinson disease. Johan Ericson and colleagues (Cell 124, 393–405; 2006) now identify Lmx1a and Msx1 as two genes that are critical in specifying midbrain dopaminergic fates in vivo. The authors found that these genes are coexpressed in dopaminergic neuronal progenitors in the ventral midbrain in mouse and chick embryos. Using a combination of overexpression and siRNA knockdown, they then demonstrated that Lmx1a is both necessary and sufficient to specify midbrain dopaminergic neurons in vivo. Furthermore, they found that Msx1 acts synergistically with Lmx1a to induce dopaminergic neurons by suppressing floor plate characteristics and promoting neuronal differentiation of ventral midline cells. Finally, the authors showed that ES cells transfected with Lmx1a and induced to differentiate as ventral neuronal progenitors yielded a much higher percentage of tyrosine hydroxylase–positive dopaminergic neurons than untransfected cells exposed to the same differentiation signals. This work identifies Lmx1a and Msx1 as intrinsic determinants of midbrain dopaminergic fate and suggests that these genes could facilitate development of stem cell–based therapies for Parkinson disease. KV

Flora of the gut

Yijuan Ruan and colleagues have taken a metagenomic approach to survey the human enteric RNA viral flora of the human gastrointestinal (GI) tract by examining uncultured RNA viruses from feces of healthy humans (PLoS Biol. 4, e3; 2006). This follows metagenomic studies characterizing DNA viruses present in human feces, many of which resembled bacteriophages. In the current study, the authors examined fecal samples from two healthy adults and constructed libraries from isolated viral RNA samples. Comparing sequences, they found that about 75% resembled viruses, and, of these, 3% resembled animal viruses, whereas 97% resembled plant viruses. The most frequent virus was pepper mild mottle virus (PMMV), a plant pathogenic virus. In order to consider how PPMV enters the human GI, the authors examined the presence of the virus in a range of pepper-based foods. They recruited three new volunteers in the same location (Southern California) and monitored their food intake before obtaining a fecal sample. Results showed that the presence of PPMV correlated to eating pepper-based foods. The authors also examined additional fecal samples from six volunteers in San Diego (three of which were positive for PPMV) and nine from Singapore (six of which tested positive), providing additional evidence to suggest this virus may be a common human guest not restricted to a single geographic location. OB

Research Highlights written by Orli Bahcall, Alan Packer and Kyle Vogan.