In situ architecture of human kinetochore-microtubule interface visualized by cryo-electron tomography

Abstract

Chromosome segregation during mitosis relies on a carefully coordinated interplay between the centromere, kinetochore, and spindle microtubules. Despite its importance, the architecture of this interface remains elusive. Here we combined in situ cryo-electron tomography and cryo-light microscopy to visualize the native architecture of the kinetochore-microtubule interface in human U2OS cells at different stages of mitosis. Our data reveal that upon microtubule binding, the centromere forms a pocket-like structure around kinetochore microtubules. This centromeric pocket contains sparsely distributed nucleosome chains and two morphologically-distinct fibrillar densities from the kinetochore that form both end-on and side-on attachments to the plus-ends of microtubules within the pocket. Our data thus suggest a dynamic kinetochore-microtubule interface with multiple interactions between outer kinetochore components and spindle microtubules.

Nanoscale analysis of human G1 and metaphase chromatinin situ

“…Ribosomes are located near the surface of the compacted chromosomes (Figure 4A), consistent with previous analysis of both HeLa cell cryosections (Eltsov et al ., 2008), U2OS cryolamellae (Zhao & Jensen, 2022), our cryosections (Figure S2), and isolated metaphase chromosomes (Nishino et al , 2012). Spindle microtubules that are inserted into metaphase chromosomes were previously observed in U2OS cells (Kamasaki et al , 2013) and are also seen in the RPE-1 chromosomes (Figure 4B).…”

“…In our cryotomograms of metaphase cells, the canonical nucleosomes are uniformly distributed within the compacted chromosomes (with the exception of pockets; see below), i.e., we did not observe separate domain-like features within the metaphase chromatin. Human metaphase chromatin has also been studied in situ in 3-D by Chrom-EMT (Ou et al ., 2017) and cryo-ET of cryo-FIB-milled cells (Zhao & Jensen, 2022). Neither of these EM-based studies reported any nanodomain-like features in mitotic chromatin.…”

“…In G1 chromatin domains, we did not find any instances of pockets, but we did find a few unambiguous ones in metaphase chromatin. With the exception of pockets, the RPE-1 metaphase chromatin distribution is uniform and resembles that seen in other cryo-EM and cryo-ET studies of mitotic human cells (Eltsov et al ., 2008; Zhao & Jensen, 2022). We did not detect mesh/network-like nucleosome-free regions analogous to the “reticular pattern” of negatively stained cells in a previous study (Ou et al ., 2017).…”

The Four Causes: The Functional Architecture of Centromeres and Kinetochores

“…Kinetochores have also been revealed to influence and organise the surrounding chromatid, establishing them as much more than machines that couple chromosomes to microtubules. Indeed, in situ cryo-electron tomography now shows how the human kinetochore is sitting within a centromeric chromatin pocket with expected filamentlike linkages to microtubules visible (274).…”

“…These k-units have been isolated from cells and can be visualised by electron microscopy (( 84 Similar counting experiments in human cells estimated ∼244 NDC80C per kinetochore (233). Recent electron tomography now indicates there may be only ~10 microtubules per human kinetochore (122,181,274). This gives ~24 NDC80C per MT in humans.…”