Tête-à-tête with CtBP dimers

Abstract

Jecrois et al. (2020) use cryoelectron microscopy to illuminate the tetrameric conformation of the CtBP2 transcriptional corepressor, a protein frequently overexpressed in human cancers. The in vivo functional characterization of tetramer-destabilizing mutants indicates that tetramerization is a physiologically important process, critical for CtBP control of gene regulation and cell migration.

The paralogous CtBP1 and CtBP2 proteins are transcriptional regulators involved in cell fate, apoptosis, and the epithelial-to-mesenchymal transition. CtBP proteins are misregulated in many human cancers including breast, colon, and ovarian cancers, leading to inhibition of apoptosis and promotion of metastasis. CtBP was first identified by the Chinnadurai lab as a cofactor binding the C terminus of the adenoviral E1A oncoprotein (Boyd et al., 1993), and has since been recognized as a transcriptional scaffold that binds histone modifiers and chromatin remodelers. Unique among transcriptional coregulators, CtBP structurally resembles D-2-hydroxyacid dehydrogenases; the protein binds NAD(H), and this has been suggested to permit the sensing of the cell’s metabolic state. Notwithstanding, the catalytic activity has remained an enigma, and an in vivo CtBP substrate has yet to be identified. Early studies focused on determining the link between the proposed dehydrogenase activity, NAD(H) binding, and transcriptional regulation by CtBP. Through biochemical and structural analysis, it was determined that NAD(H) binding leads to dimerization of CtBP, which was thought to be the physiologically relevant form of the protein (Figure 1A; Kumar et al., 2002). In conditions of low NAD(H) levels, CtBP remains monomeric; thus, a shift in NAD(H) levels may drive dimerization and activity under conditions of hypoxia. These findings provided clues to the significance of the unique features of CtBP.

Figure 1. CtBP proteins tetramerize through NAD(H) binding, and several residues at the interdimer interface are required for tetramerization.

(A) NAD(H) allows for CtBP monomers to dimerize and also form tetramers.

(B) CtBP2 cryo-EM structure shows a tetramer with five residues involved in tetramerization highlighted in the inset. The C-terminal domain, which is unstructured and not resolved, may stretch out to 280 Å (structure from Jecrois et al., 2020; PDB: 6WKW).

Further structural analysis of this important protein has revealed additional, previously unsuspected properties. Using in vitro characterization of purified recombinant protein, the Royer laboratory provided evidence of a CtBP1 and CtBP2 tetrameric state using size exclusion chromatography as well as X-ray crystallography (Bellesis et al., 2018). The biological significance of this higher-order structure, also reported by Lundblad and colleagues (Madison et al., 2013), remained unclear until now.

In this issue of Structure, Jecrois et al. (2020) use high-resolution cryoelectron microscopy (cryo-EM) to validate the important residues mediating inter-subunit contacts also seen in X-ray studies (Bellesis et al., 2018; Jecrois et al., 2020). Importantly, they exploited these molecular insights to generate point mutants that are still competent for dimerization, but abrogate tetramerization in vitro (Figure 1B). These mutants were then assessed for biological function in cell culture (Jecrois et al., 2020). Colon cancer cells null for CtBP2 (HCT116) were transfected with the five CtBP2 mutant forms and the effects on gene expression and cell migratory activity were tested. Strikingly, although the mutant proteins were expressed at similar levels to the wildtype protein, they were unable to activate TIAM1 expression or repress CHD1 expression, both cancer-related genes that are direct CtBP2 targets (Jecrois et al., 2020). These results indicate that CtBP2 mutants that are unable to tetramerize lose their transcriptional co-regulatory activities in cell culture. The mutant forms were also unable to support heightened cell migration. Taken together, these results provide strong evidence of the functional relevance of CtBP protein tetramerization.

Jecrois et al. (2020) also determined the structure of CtBP2 with its flexible C-terminal domain (CTD), although at low resolution. Unfortunately, electron density maps could not confidently assign a structure to the CTD, but the full-length protein formed a similar tetrameric structure to the truncated protein. The cryo-EM structures confirm that the CTD is not required for single-particle oligomerization. These results contrast with those of Lundblad and colleagues, who suggested that the CTD is required for tetramerization of CtBP1 in vitro (Madison et al., 2013). Yet Royer and colleagues determined that tetrameric forms of CtBP1 and CtBP2 lacking the entire C-terminal region can be obtained in vitro, although the presence of the CTD yields more tetramer formation (Bellesis et al., 2018). In fact, in some organisms, CtBP isoforms are expressed that entirely lack the CTD, which may impact tetramer formation (Mani-Telang and Arnosti, 2007). Thus, it is still unclear whether this C-terminal extension is required in vivo for stabilization of tetramers.

The five residues in CtBP2 that were found to be required for tetramerization are embedded in a larger block of ~100 amino acids that are highly conserved: three (R190, G216, and L221) are absolutely conserved across CtBP2 proteins in vertebrates, and the other two (S128 and A129) are very highly conserved (data not shown). Similar or identical residues are found in mammalian CtBP1, which can also form tetramers. In fact, the tetramer-promoting R,G,L residues are also absolutely conserved across arthropods, indicating that tetramerization may be an ancient structural property of CtBP.

This study raises additional interesting questions. Here, the CtBP2 mutants were tested in only one cellular context through overexpression; testing these tetramer-destabilizing mutants in a developmental context, through overexpression or CRISPR-mediated mutagenesis, would provide additional clues to the physiological significance of tetramerization. For instance, a catalytic domain residue, H315, was previously shown to be unnecessary in mouse embryonic fibroblasts; however, a similar genomic rescue construct containing this catalytic site mutation failed to fully rescue CtBP mutations in Drosophila (Grooteclaes et al., 2003; Zhang and Arnosti, 2011). A similar assay with the CtBP2 tetramer-destabilizing mutants in a developmental system would unequivocally determine the physiological importance of tetramerization outside of overexpression in select cell lines.

Additionally, whether tetramer assembly is a regulated and reversible process is not currently known. Early studies indicated that CtBP binds to NAD+ and NAD(H) with different affinities, leading to differences in binding to other proteins (Zhang et al., 2002; Fjeld et al., 2003), yet this difference in affinity was disputed by other studies (Kumar et al., 2002; Madison et al., 2013). Still, CtBP may function as a redox sensor, being activated through varying levels of NAD(H), impacting oligomerization of the protein and its interactions with cofactors. Regulation of tetramer assembly by endogenous processes such as post-translational modifications also remains to be determined. Regardless of whether tetramerization is normally a regulated process, the assembly of CtBP tetramers might be exploited in cancer therapeutics to control CtBP activity in certain contexts and reduce its oncogenic properties. The structural insights of CtBP will continue to reveal surprising aspects of this central player in cellular processes and disease.