Indicative of its great significance for the cell, Mps3 is essential for yeast viability. To understand which domains within the protein fulfill its essential function, a number of mps3 truncation alleles were generated and tested for their ability to support viability in an mps3 deletion strain (Fig. 20A). Being a single pass transmembrane protein of the inner nuclear membrane, the Mps3 N-terminal domain faces the nucleoplasm whereas the C-terminus protrudes into the intermembrane space. Strikingly, the SUN domain (for Mps3 domain composition please see Fig. 20B) as well as the nucleoplasmic domain is required for full viability (Fig 16A).
Results Interactors of Mps3 at the nuclear envelope
Figure 20. Mps3 nuclear- and SUN-domain are needed for full viability in yeast
(A) Viability of several mps3 truncation mutants was assayed by shuffling out the WT MPS3 upon plating on 5’FOA-containing medium. Equal amounts of cells were spotted and images were taken after 48h of growth at 30°C.
(B) Schematic representation of the Mps3 domain architecture.The N-terminus faces the nucleoplasm (aa 1-151) whereas the C-terminal part of Mps3 (aa 170-682) protrudes into the intermembrane space. The highly acidic (orange), transmembrane (TM, pink), coiled-coil (cc, green), polyglutamine (pQ, yellow) and SUN domain (blue), as well as constructs used for experiments shown in (A) and (B) are indicated.
(C) Y2H-assays. The SUN domain facilitates Mps3 homodimerization but is not needed for interaction with nuclear binding partners like Eco1. The eco1 allele ctf7-203 was used as negative control.
Sun domains are known to either mediate heterotypic interactions with KASH- domain containing proteins or homodimerization via SUN-SUN interactions. Regarding Mps3, the latter seems to be the case as Y2H constructs lacking the SUN domain no longer showed self-interaction (Fig. 20C). Deleting the SUN domain
A
C
Results Interactors of Mps3 at the nuclear envelope within Mps3 abrogates its localization to spindle pole bodies (Jaspersen et al., 2006). This is probably the reason for the observed lethality. Interestingly, this SUN- mediated dimerization is apparently not required for interaction with known nuclear binding partners such as the cohesion establishment factor Eco1 (Antoniacci et al., 2004), but also H2A.Z (Fig. 20C and data not shown). This indicates that their interaction takes place outside of spindle pole bodies. However, such interactions still seem to depend on membrane anchoring, as an Mps3 construct lacking the transmembrane domain (mps31-150) failed to interact with e.g. H2A.Z (Fig. 20C and data not shown).
Remarkably, the first 75 aa within the Mps3 N-terminus were dispensable for cell growth (Fig. 20A). Intriguingly, the remaining essential part of the nuclear domain (aa 75-150) contains the H2A.Z binding site. However, when directly testing the H2A.Z-binding defective allele (mps3∆87-93), it showed no growth defect under unperturbed growth conditions (Fig. 20A). This implies that the interaction between Mps3 and H2A.Z is in fact not essential during unchallenged cell growth and that there must be another important function executed by this essential domain.
Figure 21. The Mps3 nucleoplasmic domain is hardly required for DSB tethering
(A) Mps33myc-directed ChIP at MAT (0.2kb from DSB) in either WT or a strain lacking the Mps3
nucleoplasmic domain (mps3150-682). Shown are IP/input signals normalized to 1 for the signal before induction. Data are represented as mean ± SEM. Constructs used for ChIP are schematically shown
below the graph. Right panel: Western blot of ChIP input samples demonstrating that both WT and truncated Mps3 constructs are expressed normally.
(B) The Mps3 nucleoplasmic domain contributes to resistance to DSBs. Equal amounts of cells were spotted onto YPD plates or plates containing the DSB-inducing agent zeocin (pH 7.2). Images were taken after 48h of growth at 30°C.
Results Interactors of Mps3 at the nuclear envelope Even though interactions identified by ChIP analysis need not necessarily be direct, the fact that Mps3 gets recruited to a persistent DSB (Fig. 12) and co-purifies with DSB repair proteins (Fig. 19) suggested that Mps3 could be the actual anchor for DSB tethering at the nuclear envelope. To test this hypothesis, ChIP assays after DSB induction at MAT were performed in a strain lacking the entire Mps3 nucleoplasmic domain (mps3150-682). Surprisingly, the Mps3 nucleoplasmic domain was hardly required for DSB tethering (Fig. 21A). However, the Mps3 nucleoplasmic domain still seems to be functionally linked to DSB repair as the truncated construct rendered cells significantly sensitive to DSBs (Fig. 21B).
Given that the absence of the Mps3 nucleoplasmic domain did not significantly hamper recruitment of a persistent break to the periphery, there might be other, possibly redundant DSB-anchoring proteins at the nuclear envelope. To test this idea, ChIP assays with epitope-tagged versions of the nucleoporin Nic96 and the inner nuclear membrane protein Heh2 were performed (Fig. 22). Notably similar to Mps3, these proteins also associated with the persistent break about 3 hours after DSB induction. Intriguingly these results imply, that DSB tethering may not take place exclusively at nuclear pores nor at Mps3 or Heh2, but may involve the nuclear envelope more broadly.
Figure 22. Inner nuclear membrane protein ChIP at MAT
The integral nuclear transmembrane proteins Mps3, Heh2, and the nucleoporin Nic96 are enriched at the DSB site 4 h after HO induction, as monitored by ChIP analysis (0.2 kb from DSB). Shown are IP/input signals normalized to 1 for the signal before induction. Data are represented as mean ± SEM.
Results Possible functions of DSB relocation to the nuclear envelope