Recent findings implicate alternate core promoter recognition complexes in regulating cellular differentiation. Here we report a spatial segregation of the alternative core factor TAF3, but not canonical TFIID subunits, away from the nuclear periphery, where the key myogenic gene MyoD is preferentially localized in myoblasts. This segregation is correlated with the differential occupancy of TAF3 versus TFIID at the MyoD promoter. Loss of this segregation by modulating either the intranuclear location of the MyoD gene or TAF3 protein leads to altered TAF3 occupancy at the MyoD promoter. Intriguingly, in differentiated myotubes, the MyoD gene is repositioned to the nuclear interior, where TAF3 resides. The specific high-affinity recognition of H3K4Me3 by the TAF3 PHD (plant homeodomain) finger appears to be required for the sequestration of TAF3 to the nuclear interior. We suggest that intranuclear sequestration of core transcription components and their target genes provides an additional mechanism for promoter selectivity during differentiation.

Commentary: Jie Yao in Bob Tijan’s lab used a combination of confocal microscopy and dual label PALM in thin sections cut from resin-embedded cells to show that certain core transcription components and their target genes are spatially segregated in myoblasts, but not in differentiated myotubes, suggesting that such spatial segregation may play a role in guiding cellular differentiation.

The challenge of recovering the topology of massive neuronal circuits can potentially be met by high throughput Electron Microscopy (EM) imagery. Segmenting a 3-dimensional stack of EM images into the individual neurons is difficult, due to the low depth-resolution in existing high-throughput EM technology, such as serial section Transmission EM (ssTEM). In this paper we propose methods for detecting the high resolution locations of membranes from low depth-resolution images. We approach this problem using both a method that learns a discriminative, over-complete dictionary and a kernel SVM. We test this approach on tomographic sections produced in simulations from high resolution Focused Ion Beam (FIB) images and on low depth-resolution images acquired with ssTEM and evaluate our results by comparing it to manual labeling of this data.