Smad proteins are key components of the TGF-β pathway, one of the four conserved pathways in metazoans. In humans, these proteins are involved in a wide range of cellular responses and are altered in many oncologic and autoimmune processes. Through our research, we have defined a phosphorylation-dependent mechanism that labels Smads, first for activation and then for degradation, and shown how transcription cofactors and E3 ubiquitin ligases interact with Smad proteins. We have also clarified the “auto-ubiquitination process” of Nedd4L, which, in fact, requires two ubiquitin ligases, one of them having a partially unfolded catalytic domain that becomes the target for a fully functional ligase. A second topic of our research has aimed at discovering how WW domains recognize specific targets and how domain plasticity correlates with misfolding processes and aggregation.

The gene responses activated by the TGF-β cytokine family play important roles in embryo development, apoptosis, tissue homeostasis, repair, and immunity. These critical roles demand a high level of conservation and fidelity of the TGF-β signaling elements in healthy organisms. The main TGF-β signal transduction mechanism is the Smad pathway. The formation of large nuclear assemblies of Smads with chromatin remodelers and with master transcription factors directs these molecular machines to binding sites of promoters and enhancer regions, tuning their function to specific cellular contexts. Mutations in these key components of TGF-β signaling are responsible for various inherited and somatic diseases in humans.

The modular composition of Smad proteins includes an N-terminal MH1 domain, which binds to DNA, a linker, and the C-terminal MH2 domain, which enables the interaction with other Smads as well as with transcription factors and repressors (TiBS 2015). The MH1 domain is exclusive of Smad proteins. However, MH2 sequences are also found in non-Smad proteins present only in deuterostomes (invertebrates), which do not participate in TGF-β signaling. Using X-ray crystallography, a technique we incorporated to our lab’s expertise in 2014, we investigated the reason for this, and determined the first structure of a non-Smad MH2 domain. We found that, although the structure displays the main features of the MH2 fold, the protein contains an additional α-helical region blocking an interface of the domain that in the canonical Smad MH2 domains serves to interact with other Smads, thus explaining why this protein cannot participate in TGF-β signaling (Acta Crystallographica Section D, 2015).

The commonly accepted paradigm suggested that the TGF-β-activated Smads (Smad2/3) and Smad4 showed a preference for the GTCT site (known as the Smad Binding Element, SBE), whereas the BMP-activated Smads (Smad1/5/8) preferred GC-rich motifs. However, the sequence conservation of the MH1 domains, together with recent experimental evidences in the literature and our own work, indicate that the separation between DNA binding preferences of R-Smads is subtler than initially thought. Our structures revealed that the MH1 domains of Smad3 and Smad4 proteins interact efficiently and specifically with GC-rich motifs. We classified these motifs in the 5GC consensus GGC(GC)|(CG). Together with the group of Dr. J. Massagué we found that these motifs were functionally relevant for TGF-β-activated Smads and for Smad4 in the goosecoid promoter, a nodal-regulated differentiation gene, (Nature Communications, 2017).

We also found that all MH1 domains bind SBE and 5GC sites using a similar binding mode, although Smad2/3/4 MH1 domains bind to the DNA as monomers, whereas Smad1/5/8 form N-terminal helix-swapped dimers. We found that dimer/monomer preferences correlate with the loop1 length and sequence, since swapping the loop1 between Smad5 and Smad3 results in the Smad5 chimera-DNA complex crystallizing as a monomer. In fact, in eukaryotes, the association between transcription factors (TFs) to form homo- and hetero-dimers is a common feature and it seems that BMP-activated Smads follow this rule. These dimers can be formed in the full-length protein context, facilitated by the flexibility provided by the long linkers (80 residues) connecting the MH1 and MH2 domains. Once we clarified that monomers or dimers are intrinsic properties of MH1 domains, we explored their implications for Smad function. Combining our information with that available in the literature, we propose how dimers and monomers of MH1 domains can influence the composition of Smad full-length proteins and DNA binding. This work also helps interpret previously unexplained observations on Smad complexes interacting with so many promoters and enhancers.

We have also investigated how signal-activated Smad transcription factors cooperate with repressors. We first studied how TGIF1–Homeodomain (TGIF1–HD) binds to its canonical DNA motif. We have also shown that this transcription factor directly interacts with the MH1 domain of Smad proteins. Moreover, the formation of the HD-MH1 complex partially hinders the DNA-binding site of the complex, preventing the efficient interaction of TGIF1–HD and Smads with DNA, thus revealing how the Smad-TGIF1 complex acts as a transcriptional repression system (NAR 2018).

We also studied how Smad2/3 interact with pioneer factors (FoxH1) to activate differentiation. Together with the group of Dr. J. Massagué́ we studied TGF-β/Nodal activation of mesendoderm differentiation in pluripotent progenitors during mammalian gastrulation. Our structures revealed that the functional differences between Smad2 and Smad3 and their interactions with DNA lie within a small segment of Smad2 MH1 domain, absent in Smad3, which populates a series of conformations that regulate access of the DNA-binding hairpin to the DNA. Since only some conformations can interact with DNA, it turned out that Smad2 is less competitive than Smad3 at binding to target promoters (Genes and Development, 2019).

Highlights

• We have discovered a mechanism that labels these key components of the TGF-beta pathway first for activation and then for degradation.
• Smads accumulate tumor mutations in the binding interfaces with other Smads, DNA and cofactors.
• We have defined specific 5GC DNA motifs that interact with R-Smads and with Smad4.
• Functional differences of TGFβ- and BMP-activated R-Smads are not related to their DNA specificity. Every R-Smad and Smad4 can interact with SBE and with new 5GC sites using a conserved binding mode.
• We propose that dimer/monomer propensities of MH1 domains help define the composition of Smad complexes and dictate how Smads interact with cis-regulatory elements.
• Smad2 and Smad3 have different functions. Well folded Smad2 binds DNA. This result challenges an erroneous view that had prevailed in the field for two decades, that Smad2 does not interact with DNA.

Relevant publications:

1. Macias, M.J.*; Martin-Malpartida, P. and Massagué J.* (review). Structural determinants of Smad function in TGF-beta signaling. TiBS, 40(6), 296-308, 2015.
2. Beich-Frandsen M, Aragón E, Llimargas M, Benach J, Riera A, Pous J, Macias M.J.* Structure of the N-terminal domain of the protein Expansion: an 'Expansion' to the Smad MH2 fold. Acta Crystallogr D Biol. 2015.
3. (Martin-Malpartida P., Batet M., Kaczmarska Z.), Freier R., Gomes T., Aragón E., Zou Y., Wang Q., Xi Q., Ruiz L., Vea A., Márquez J.A., Massagué J., and Macias M.J.* Structural basis for genome wide recognition of 5-bp GC motifs by SMAD transcription factors. Nat. Commun. 8(1), 2070, 2017.
4. Guca E., Suñol D., Ruiz L., Konkol A., Cordero J., Torner C., Aragon E., Martin-Malpartida P., Riera A., Macias M.J.* TGIF1 homeodomain interacts with Smad MH1 domain and represses TGF-β signaling. NAR 46(17), 9220- 9235, 2018.
5. (Aragón E., Wang Q., Zou Y.), Morgani S.M., Ruiz L., Kaczmarska Z., Su J., Torner C., Tian L., Hu J., Shu W., Agrawal S., Gomes T., Márquez J.A., Hadjantonakis A.K., Macias M.J.* and Massagué J.* Structural basis for distinct roles of SMAD2 and SMAD3 in FOXH1 pioneer-directed TGF-β signaling. Genes & Development, 2019
6. (Ruiz L., Kaczmarska Z., Gomes T.), Aragon E., Torner C., Freier R., Baginski B., Martin-Malpartida P., de Martin Garrido N., Marquez J.A., Cordeiro T., Pluta R., Macias M.J.* Unveiling the dimer/monomer propensities of Smad MH1-DNA complexes (under review, 2019) 10.1101/833319 on bioRxiv