Gaudin (page 601) used a liposome assay to determine the sensitivity of rabies virus–induced membrane fusion to lysophosphatidylcholine (LPC) and oleic acid. The effects of these lipids on the rabies virus–induced fusion pathway are similar to their effects in other fusion systems, supporting the idea that all biological membrane fusion events proceed by the same pathway.

Although extensive work on influenza virus–induced membrane fusion and theoretical models of the energetics of lipid structures have led to the proposal that membrane fusion proceeds by a universal pathway, this theory has only been tested in a small number of systems. In the new work, the author focused on rabies virus–induced membrane fusion, which differs from influenza virus–induced fusion in several respects. The response of the rabies virus fusion event to LPCs with alphatic side chains of varying length, and to oleic acid, demonstrates that rabies virus and influenza virus–induced fusion follow similar pathways.

In both cases, fusion begins with the insertion of a fusion peptide into the target membrane, which then induces the formation of a stalk structure connecting the viral and target membranes. The stalk then expands, and the inner leaflets of the membranes form a hemifusion diaphragm, which then forms a pore that expands and completes the fusion process. The ability to make direct comparisons between structurally diverse membrane fusion systems should facilitate higher resolution studies of membrane dynamics.

Using a combination of genetic crosses and cell culture assays, Gayraud et al. (page 667) have developed a new model to explain the phenotype of Tight skin (Tsk) mutant mice, which carry a mutation in the fibrillin 1 (Fbn1) gene. In addition to providing a new perspective on the assembly of extracellular microfibrils, the data help to explain the discrepancies between Tsk mutant mice and Marfan syndrome, a human genetic disorder. Like Marfan syndrome patients, mice with reduced Fbn1 gene expression exhibit lung emphysema, bone overgrowth, and vascular complications. Heterozygous Tsk/+ mice, which express a mutant Fbn1 gene product at normal levels, display lung emphysema and bone overgrowth, but lack the vascular complications typical of Marfan syndrome and the other mouse models.

In the new work, the authors demonstrate that Tsk fibrillin 1 copolymerizes with wild-type fibrillin 1, in contrast to earlier models which suggested the formation of two distinct pools of microfibrils in Tsk/+ mice. Instead, Tsk fibrillin 1 molecules appear to participate in the initial stages of microfibril assembly, but can only yield long microfibrils by copolymerizing with wild-type fibrillin 1. The results suggest that copolymerization produces functionally deficient microfibrils, decreasing the level of functional microfibrils below the threshold necessary to cause Marfan syndrome-like bone overgrowth and emphysema, but not below the threshold for vascular abnormalities.

Beginning on page 433, Dundr et al. describe the first examination of the timing of postmitotic nucleolar reassembly using GFP-tagged proteins in living cells. The results support earlier findings that postmitotic nucleoli incorporate assembled components derived from the maternal cell, and also provide temporal data that suggest a detailed model of nucleolar assembly.

By following the localization of GFP fusions of the processing-related proteins fibrillarin, nucleolin, and B23, they found that the nucleolus-derived foci (NDF), which contain partially processed preribosomal RNA in association with processing components, disappear during telophase. At the same time, the tagged proteins begin to appear in the reforming nuclei of the daughter cells. Prenucleolar bodies (PNBs), which appear in nuclei in early telophase, gradually disappear as the nucleoli form, suggesting that the PNB components are transferred to the forming nucleoli.

Based on their results, the authors propose a model in which pre-rRNA transcripts, associated with processing components, are transferred to postmitotic nuclei, where they become incorporated into PNBs. The processing components are then transferred from PNBs into the newly forming nucleoli in a process that depends on the reactivation of nucleolar transcription. At the same time, reactivation of pre-rRNA processing also seems to occur. The team is now hoping to use the GFP-tagged components to confirm the details of the new model and to elucidate the regulatory mechanisms controlling nucleolar reassembly.

Although several groups have analyzed the structure of tRNA bound to the ribosome by cryo-electron microscopy (cryo-EM) and x-ray crystallography, the functionally relevant pre- and posttranslocation tRNA binding sites on the ribosome have remained uncertain. Agrawal et al. (page 447) now report the development of cryo-EM maps of tRNA–ribosome complexes at a resolution of 17Å under precisely controlled conditions, revealing the principal positions of tRNA through the course of the elongation cycle.

The new results constitute the highest resolution three-dimensional cryo-EM maps of tRNA-ribosome complexes to date, and the carefully controlled buffer conditions address a problem discovered in earlier work: tRNA positions can vary significantly in different buffers. The team identified six different tRNA locations, named A (before peptide bond formation), APep (after bond formation), P, P/E, E, and E2. Although some of the site assignments agree with those previously described, Agrawal et al. also identified several differences between their results and those of earlier x-ray and cryo-EM studies. Based on the accumulated structural data, the authors present a detailed model of the main events of the elongation cycle.

By examining the localization of RNA transcribed from mutant and wild-type COL1A1 genes in heterozygous cells, Johnson et al. (page 417) found that while mutant transcripts initiate transport from the gene, they are unable to exit the SC-35 domain adjacent to the gene. In addition to providing new information about the molecular pathogenesis of osteogenesis imperfecta, the genetic disease associated with COL1A1 mutations, the new results identify a previously unknown step in mRNA export and may help resolve the controversy over the existence of RNA “tracks” in the nucleus.

Nuclear RNA “tracks,” localized accumulations of post-transcriptional RNA, appear beside both the normal and mutant COL1A1 genes. The normal COL1A1 RNA tracks are distributed throughout a large SC-35-containing domain on one side of the gene, and the transcripts are dispersed from the domain after splicing is complete. Mutant transcripts initiate transport from the gene, but are then retained within the SC-35 domain, identifying an early step in the mRNA export pathway.

The results suggest that normal COL1A1 mRNA generates a “track” by moving from the gene into the adjacent SC-35 domain, then disperses from the domain into the nucleoplasm before reaching the nuclear envelope and being translocated into the cytoplasm. Surprisingly, the COL1A1 transcripts are predominantly spliced before they passage through the splicing factor domain, where they find evidence that further maturation and release for export occur. The authors suggest that SC-35 domains associated with other genes may function similarly.

Alan W. Dove, 350 E. Willow Grove #406, Philadelphia, PA, 19118. E-mail: