Developmental regulation and spatiotemporal redistribution of the sumoylation machinery in the rat central nervous system

Background

Small Ubiquitin-like MOdifier protein (SUMO) is a key regulator of nuclear functions but little is known regarding the role of the post-translational modification sumoylation outside of the nucleus, particularly in the Central Nervous System (CNS).

Methodology/Principal Findings

Here, we report that the expression levels of SUMO-modified substrates as well as the components of the sumoylation machinery are temporally and spatially regulated in the developing rat brain. Interestingly, while the overall sumoylation is decreasing during brain development, there are progressively more SUMO substrates localized at synapses. This increase is correlated with a differential redistribution of the sumoylation machinery into dendritic spines during neuronal maturation.

Conclusions/Significance

Overall, our data clearly demonstrate that the sumoylation process is developmentally regulated in the brain with high levels of nuclear sumoylation early in the development suggesting a role for this post-translational modification during the synaptogenesis period and a redistribution of the SUMO system towards dendritic spines at a later developmental stage to modulate synaptic protein function.     

A Role for Non-Covalent SUMO Interaction Motifs in Pc2/CBX4 E3 Activity

Background

Modification of proteins by the small ubiquitin like modifier (SUMO) is an essential process in mammalian cells. SUMO is covalently attached to lysines in target proteins via an enzymatic cascade which consists of E1 and E2, SUMO activating and conjugating enzymes. There is also a variable requirement for non-enzymatic E3 adapter like proteins, which can increase the efficiency and specificity of the sumoylation process. In addition to covalent attachment of SUMO to target proteins, specific non-covalent SUMO interaction motifs (SIMs) that are generally short hydrophobic peptide motifs have been identified.

Methodology/Principal Findings

Intriguingly, consensus SIMs are present in most SUMO E3s, including the polycomb protein, Pc2/Cbx4. However, a role for SIMs in SUMO E3 activity remains to be shown. We show that Pc2 contains two functional SIMs, both of which contribute to full E3 activity in mammalian cells, and are also required for sumoylation of Pc2 itself. Pc2 forms distinct sub-nuclear foci, termed polycomb bodies, and can recruit partner proteins, such as the corepressor CtBP. We demonstrate that mutation of the SIMs in Pc2 prevents Pc2-dependent CtBP sumoylation, and decreases enrichment of SUMO1 and SUMO2 at polycomb foci. Furthermore, mutational analysis of both SUMO1 and SUMO2 reveals that the SIM-interacting residues of both SUMO isoforms are required for Pc2-mediated sumoylation and localization to polycomb foci.

Conclusions/Significance

This work provides the first clear evidence for a role for SIMs in SUMO E3 activity.     

Yeast PIAS-type Ull1/Siz1 is composed of SUMO ligase and regulatory domains

SUMO (small ubiquitin-like modifier)/Smt3 (suppressor of mif two) is a member of the ubiquitin-related protein family and is known to conjugate with many proteins. In the sumoylation pathway, SUMO/Smt3 is transferred to substrate lysine residues through the thioester cascade of E1 (activating enzyme) and E2 (conjugating enzyme), and E3 (SUMO ligase) functions as an adaptor between E2 and each substrate. Yeast Ull1 (ubiquitin-like protein ligase 1)/Siz1, a PIAS (protein inhibitor of activated STAT)-type SUMO ligase, modifies both cytoplasmic and nuclear proteins. In this paper, we performed a domain analysis of Ull1/Siz1 by constructing various deletion mutants. A novel conserved N-terminal domain, called PINIT, as well as the RING-like domain (SP-RING) were required for the SUMO ligase activity in the in vitro conjugation system and for interaction with Smt3 in an in vitro binding assay. The most distal N-terminal region, which contains a putative DNA-binding SAF-A/B, Acinus, and PIAS (SAP) motif, was not required for the ligase activity but was involved in nuclear localization. A strong SUMO-binding motif was identified, which interacted with Smt3 in the two-hybrid system but was not necessary for the ligase activity. The most distal C-terminal domain was important for stable localization at the bud neck region and thereby for the substrate recognition of septins. Furthermore, the C-terminal half conferred protein instability on Ull1/Siz1. Taken together, we conclude that the SP-RING and PINIT of Ull1/Siz1 are core domains of the SUMO ligase, and the other domains are regulatory for protein stability and subcellular localization.     

Functional analyses of vertebrate TCF proteins in C. elegans embryos

In the canonical Wnt pathway, signaling results in the stabilization and increased levels of β-catenin in responding cells. β-catenin then enters the nucleus, functioning as a coactivator for the Wnt effector, TCF/LEF protein. In the absence of Wnt signaling, TCF is complexed with corepressors, together repressing Wnt target genes. In C. elegans, Wnt signaling specifies the E blastomere to become the endoderm precursor. Activation of endoderm genes in E requires not only an increase in β-catenin level, but a concomitant decrease in the nuclear level of POP-1, the sole C. elegans TCF. A decrease in nuclear POP-1 levels requires Wnt-induced phosphorylation of POP-1 and 14-3-3 protein-mediated nuclear export. Nuclear POP-1 levels remain high in the sister cell of E, MS, where POP-1 represses the expression of endoderm genes. Here we express three vertebrate TCF proteins (human TCF4, mouse LEF1 and Xenopus TCF3) in C. elegans embryos and compare their localization, repression and activation functions to POP-1. All three TCFs are localized to the nucleus in C. elegans embryos, but none undergoes Wnt-induced nuclear export. Although unable to undergo Wnt-induced nuclear export, human TCF4, but not mouse LEF1 or Xenopus TCF3, can repress endoderm genes in MS, in a manner very similar to POP-1. This repressive activity requires that human TCF4 recognizes specific promoter sequences upstream of endoderm genes and interacts with C. elegans corepressors. Domain swapping identified two regions of POP-1 that are sufficient to confer nuclear asymmetry to human TCF4 when swapped with its corresponding domains. Despite undergoing Wnt-induced nuclear export, the human TCF4/POP-1 chimeric protein continues to function as a repressor for endoderm genes in E, a result we attribute to the inability of hTCF4 to bind to C. elegans β-catenin. Our results reveal a higher degree of species specificity among TCF proteins for coactivator interactions than for corepressor interactions, and uncover a basic difference between how POP-1 and human TCF4 steady state nuclear levels are regulated.