Antiviral mammalian microRNA

RNA silencing is known to have a defensive role against viral infection in plants and insects, and Olivier Voinnet and colleagues now show that RNA silencing also has an analogous function in mammalian cells (Science 308, 557–560; 2005). When searching for small RNAs derived from primate foamy virus type I (PFV-1) in infected 293T cells, they instead identified a cellular microRNA (miRNA) that limits PFV-1 replication. Human miR-32 targets PFV-1 open reading frame 2 and exerts a sequence-specific effect to restrict virus production. Voinnet and colleagues also show that PFV-1 counters this cellular defense with a silencing suppressor, encoded by the viral gene tas. The Tas protein, already known to be a transcriptional transactivator, promotes the nonspecific accumulation of cellular miRNAs and thus functions as a general suppressor of RNA silencing. Tas can suppress RNA silencing in Arabidopsis thaliana, suggesting that it targets a conserved step in the RNA silencing pathway. Though it does not answer the question of whether mammalian cells generate viral-derived small RNAs to counter infection, this work indicates yet another important function for miRNAs encoded in the human genome. EN

miRNA-directed phasing

miRNAs act as post-transcriptional regulators of gene expression in plants and animals by binding to complementary mRNA sequences and targeting these transcripts for either degradation or translational repression. Now, Jim Carrington and colleagues (Cell 121, 207–221; 2005) describe a new function for miRNAs in plants. Combining computational searches with functional assays, they show that two miRNAs, miR173 and miR390, function in the biogenesis of trans-acting short interfering RNAs (siRNAs), a second class of small regulatory RNAs that arise through directed cleavage of long, primary double-stranded RNA transcripts. Carrington and colleagues show that miR173 and miR390 recognize complementary sites in several trans-acting siRNA transcripts and direct the phasing of primary transcript cleavage. Using a heterologous system, they show that miR173 is necessary and sufficient to direct cleavage of primary transcripts produced by the trans-acting siRNA TAS1 and TAS2 loci. They further show that this directed cleavage depends on the presence of complementary sequences in the target transcripts. This work identifies a new role for miRNAs as positive effectors of trans-acting siRNA biogenesis, in addition to their well-defined roles as negative regulators of target gene expression. KV

miRNAs and auxin homeostasis

The ARGONAUTE1 protein of Arabidopsis thaliana is the founding member of a family of proteins involved in cleavage of miRNA-targeted mRNAs. Three recent studies have now shown AGO1 to be a key regulator of auxin homeostasis. Auxin is a multifunctional and context-dependent phytohormone that controls cell growth, cell elongation, morphogenesis and patterning in A. thaliana. Céline Sorin and colleagues explain a failure of adventitious root formation in ago1 mutants by identifying an upregulation of ARF17, a transcriptional repressor in the auxin signaling pathway that is normally targeted for degradation by miRNAs (Plant Cell 17, 1343–1359; 2005). Likewise, Allison Mallory and colleagues show that a miRNA-resistant version of ARF17 leads to increased levels of the mRNA and altered expression of ARF17 targets. This altered expression correlates with many of the previously reported phenotypes of plants with mutations in genes involved in miRNA synthesis (Plant Cell 17, 1360–1375; 2005). Finally, Hui-Shan Guo and colleagues show that miR164 guides the cleavage of NAC1, which encodes a molecule that transduces auxin signals for lateral root emergence (Plant Cell 17, 1376–1386; 2005). All told, miRNA regulation provides a mechanism for the fine-tuning of auxin signaling in many aspects of plant development. AP

Noisy networks

Studies of engineered cells have yielded considerable insight into the stochastic elements of gene regulation. Some of these studies, however, have been limited by examining populations of cells. In a recent study, Michael Elowitz and colleagues examine the relationship between the concentration of active transcription factor and the rate of protein produced in single cells (Science 307, 1962–1965; 2005). The gene regulation function (GRF), which models this relationship, had previously been defined using estimates based on single measurements of steady-state concentrations of transcription factor and protein. In the current study, the authors construct a synthetic regulatory cascade in E. coli that allows for simultaneous and dynamic measurement of these input and output signals with time-lapse microscopy. The single cell measurements of the GRF were significantly different than the population average, and the individual cell variation was largely attributable to extrinsic (fluctuations affecting global gene expression) rather than intrinsic (fluctuations specific to the individual gene) noise. These studies bring a welcome degree of quantification to the field. It will be of interest to see how they may assist in studying cellular networks, and how the principles for noise propagation in a network may be generalized when tested with a range of endogenous regulatory pathways. OB

Departure from HWE

Hardy-Weinberg equilibrium (HWE), given by the formula p2 + 2pq +q2 = 1, uses allele frequencies to determine expected genotype frequencies in a randomly mating population. That a marker adheres to HWE is often used as a test to identify genotyping error in population-based data sets to which checks for mendelian consistency cannot be applied. But departure from HWE (DHW) in an association study can indicate that the test locus is close to the disease susceptibility locus. Nancy Cox and colleagues now report an analytic framework designed to help investigators determine whether DHW is due to disease susceptibility or to genotyping errors, chance or failure to meet the assumptions of HWE (Am. J. Hum. Genet. 76, 967–986; 2005). They define Δ, the difference between observed and expected genotypic frequencies, for a number of genetic disease models and propose the use of a goodness-of-fit test to determine whether DHW is consistent with the best-fit disease model. The authors propose that examination of HWE has been underused as a tool for the identification of susceptibility loci; this conclusion will hopefully discourage the wholesale discard of markers that depart from HWE in favor of a more rational approach to investigate the underlying biology. EN

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