Filling the gap: Micro-C accesses the nucleosomal fiber at 100–1000 bp resolution

In addition to investigating new links between fine chromosomal structures and transcription, the Micro-C assay gave the authors the opportunity to assess existing models of the yeast nucleosomal fiber. On the basis of the relatively short linker length between consecutive nucleosomes in yeast (20 bp) two alternative structures for the nucleosomal fiber have been proposed (see, for example, [9] for a review; Fig. 1b). Consecutive nucleosomes (n and n + 1) can be stacked upon each other, resulting in a columnar arrangement that has been proposed to be further wrapped into a solenoidal structure [10]. Alternatively, the closest neighbors in space can also correspond to nucleosomes that occupy every two positions linearly (n and n + 2), resulting in a zig-zag motif that can be further stabilized by nucleosome-stacking interactions. Surprisingly, the inter-nucleosomal contacts reported by Hsieh et al. are compatible with both models as the number of (n/n + 1) contacts is roughly similar to the number of (n/n + 2) contacts (schematized in Fig. 1c, but see Figure S3 of Hsieh et al. [5]). These findings suggest either a dynamic equilibrium between these two structures or the absence of a highly structured nucleosomal fiber. The lack of any periodicity at 4–6 nucleosome spacing strongly suggests that the columnar phase, if it exists, is not wrapped into any higher-order periodical structure, as was proposed in pioneering studies on the chromatin fiber structure [10]. In addition, it could be argued that the asynchronous populations used to perform the experiment contain diverse structures that correlate with the diverse stages of the cell cycle. Therefore, it may be interesting to perform Micro-C on synchronized cells to search for such effects.

In line with the possibility of a polymorphic structure, Hsieh et al. show that several factors can change what they describe as the ‘compaction’ of the chromosomal fiber. Here, the compaction is simply defined as the ratio of long-range over short-range contacts (with short-range being defined as closer than 300 bp). The compaction of each gene was found to be correlated with its transcriptional activity, and the decrease in compaction observed for highly transcribed genes might be attributed to the local disruption of the nucleosomal fiber by active RNA polymerase(s). Consistent with this finding, genes that were upregulated following a diamide treatment were convincingly shown to be less compacted.

To investigate the mechanistic basis of gene compaction further, the authors set out to repeat the Micro-C experiment in 23 S. cerevisiae mutants in which chromatin structure is altered. Micro-C confirmed the role of the RSC chromatin remodeling complex and the cohesin complex in chromatin structuring, with defects in these complexes being associated with increased gene compaction. Conversely, other chromatin mutants, such as those with defects in Mediator or the histone deacetylase Rpd3, appeared to induce a significantly reduced level of gene compaction. This change is, however, accompanied by only small changes in transcriptional activity. The effect of these ‘structural mutants’ suggests that changes in compaction can also be modulated by factors other than the progression of DNA polymerase along the genes, either by the potential formation of transient long-range loops or by the modification of inter-nucleosomal interactions. To test for the latter specifically, the authors used variants of histone H4 that were previously characterized in vitro as having a direct effect on nucleosome array folding. Micro-C confirmed the important role of the H4 N-terminal tail in vivo.

Finally, the authors took advantage of their high-resolution assay to search for the promoter and terminator gene loops identified previously using a 3C approach [3]. Quite unexpectedly, such structures were not apparent in their data. However, the transcription regulator protein Ssu72, reported to be essential for the loop formation, resulted in a small but significant reduction in gene compaction. These results suggest that the reports of loops may need to be looked at again through more globular or chromatin-accessibility focused lenses.