Science Spotlight

C vs T: distinct or intertwined pathways to kinetochore assembly?

Model of the protein architecture at the human centromere based on the results of this study. CENP-T lies between two adjacent CENP-A nucleosomes and is a part of the overall CENP-A/CENP-C/CENP-B complex that assembles on "young" (closer to the center) centromeric alpha satellite repeats.
Image based on a graphic in the publication.

The DNA from a single cell in your body measures roughly six feet in length if stretched out from end to end.  To fit inside the cell, DNA is wound around protein complexes called nucleosomes and then folded even tighter into structures called chromosomes.  When cells reproduce, the DNA is replicated and one copy of each chromosome is inherited by the daughter cells.  Centromeres are the places on each chromosome that are connected to spindle fibers to pull them into the daughter cells during cell division (reproduction). Despite their functional importance, the exact DNA sequence and protein complexes present at human centromeres have remained mysterious. 

The DNA sequence of centromeres is highly repetitive, composed of sections called alpha satellites that are difficult to piece together into a continuous sequence due to their repetitive nature. There are also two different types of nucleosomes, a canonical nucleosome which contains histone 3 (H3) protein and a centromeric variant where centromere protein A (CENP-A) replaces H3.  CENP-A defines the base of the large protein complex called the kinetochore that forms over the centromeric DNA and physically connects it to spindle fibers during cell division. CENP-A physically associates with centromere protein C (CENP-C), which in turn associates with a larger network of proteins called KMN (for its constituents the KNL1, Mis12, and Ndc80 subcomplexes).  The KMN network, and the Ndc80 subcomplex in particular, directly binds spindle fibers. 

Centromere protein T (CENP-T) has recently been shown to provide a so-called "alternate" pathway of kinetochore assembly, linking the centromere to KMN.  Part of the CENP-T protein is structurally similar to a histone protein and it can associate with DNA in cells and in vitro.  On its other end, CENP-T protein binds directly to the Ndc80 complex.  The spatial relationship between CENP-C and CENP-T at the centromere has been controversial due to conflicting data from several different studies.  In their recent publication in Genome Research, scientists in the Henikoff Laboratory (Basic Sciences Division) used chromatin-immunopreciptation (ChIP), deep sequencing, and sequential ChIP to uncover that CENP-T, CENP-C, CENP-B, and two CENP-A nucleosomes form a single complex at human centromeric alpha satellites.

The scientists began by using a Micrococcal nuclease (MNase)-based native chromatin immunoprecipitation (ChIP) to isolate CENP-C or CENP-T and then analyze the associated DNA. In this method, MNase digests the "unprotected" DNA which is not bound to proteins and an antibody is used to purify protein-DNA complexes (chromatin) under cell-like ("native") conditions.  Using this protocol, they detected enrichment of CENP-C at centromeric alpha-satellite sequences as expected but did not detect enrichment of CENP-T at centromere alpha satellites.  However, the scientists knew that the lack of detection of CENP-T at the centromere could be due to a difficulty in purifying CENP-T versus CENP-C.  After trying various purification strategies, they uncovered that CENP-T and CENP-C were both found enriched on centromeric alpha satellite DNA when a lower concentration of MNase was used.  When they performed non-native ChIP by crosslinking DNA and protein together using formaldehyde, they also detected enrichment of CENP-T as well as CENP-C at centromere alpha satellites.  Therefore, the lack of CENP-T detected at the centromere by high-MNase native ChIP strategies was due to the MNase treatment and this may explain the previous results that did not detect an association of CENP-T with CENP-A nucleosomes.

When they analyzed the DNA associated with CENP-C, CENP-T, or CENP-A from the crosslinking ChIP experiments, they found that each protein protected a similar length of alpha satellite DNA, roughly 340 base pairs (bp). They hypothesized that CENP-A, CENP-C, and CENP-T may form a complex.  To test this, they performed sequential ChIP purifications under crosslinking conditions to strengthen protein-protein and protein-DNA interactions.  Starting with purified CENP-A chromatin (CENP-A ChIP), they then performed a secondary immunoprecipitation against either CENP-C or CENP-T to test whether these proteins were associating with CENP-A in the chromatin.  Interestingly, both CENP-C and CENP-T ChIP following CENP-A ChIP led to a significant enrichment in alpha-satellite centromere sequences.  Along with existing data in the field supporting the fact that CENP-C and CENP-B bind CENP-A nucleosomes, their results suggest that CENP-C, CENP-T, CENP-B, and CENP-A form a complex at human centromeres.  Further investigation uncovered that CENP-T protects the linker sequence between the two CENP-A nucleosomes that are positioned over a single alpha-satellite repeat (see figure).

"Without relying on linear maps, we have successfully mapped multiple centromeric proteins on unassembled alpha satellites using the genomic readout of in vivo biochemistry" said lead author and postdoctoral fellow Dr. Jitendra Thakur.  Additionally, the study provides evidence that the CENP-C and CENP-T pathways to kinetochore assembly are interconnected.
"The more that we learn about the molecular architecture of the centromere complex, the more glaring is our ignorance of the genomic context in which the centromere is embedded," said principal investigator Dr. Steve Henikoff.  "The human genome sequence was declared finished in 2001, while still leaving huge gaps spanning all of our centromeres. However, advances in DNA sequencing technologies now give us hope that we will soon fill in these gaps and so gain a fuller understanding of centromere function and evolution."

Thakur J, Henikoff S.  2016.  "CENP-T bridges adjacent CENP-A nucleosomes on young human alpha satellite dimers." Genome ResearchEpub ahead of print.

This research was funded by Howard Hughes Medical Institute.