Insights into Strand Exchange in BTB Domain Dimers from the Crystal Structures of FAZF and Miz1

https://doi.org/10.1016/j.jmb.2010.05.028Get rights and content

Abstract

The BTB domain is a widely distributed protein–protein interaction motif that is often found at the N-terminus of zinc finger transcription factors. Previous crystal structures of BTB domains have revealed tightly interwound homodimers, with the N-terminus from one chain forming a two-stranded anti-parallel β-sheet with a strand from the other chain. We have solved the crystal structures of the BTB domains from Fanconi anemia zinc finger (FAZF) and Miz1 (Myc-interacting zinc finger 1) to resolutions of 2.0 Å and 2.6 Å, respectively. Unlike previous examples of BTB domain structures, the FAZF BTB domain is a nonswapped dimer, with each N-terminal β-strand associated with its own chain. As a result, the dimerization interface in the FAZF BTB domain is about half as large as in the domain-swapped dimers. The Miz1 BTB domain resembles a typical swapped BTB dimer, although it has a shorter N-terminus that is not able to form the interchain sheet. Using cysteine cross-linking, we confirmed that the promyelocytic leukemia zinc finger (PLZF) BTB dimer is strand exchanged in solution, while the FAZF BTB dimer is not. A phylogenic tree of the BTB fold based on both sequence and structural features shows that the common ancestor of the BTB domain in BTB-ZF (bric à brac, tramtrack, broad-complex zinc finger) proteins was a domain-swapped dimer. The differences in the N-termini seen in the FAZF and Miz1 BTB domains appear to be more recent developments in the structural evolution of the domain.

Introduction

There are 43 BTB-ZF (bric à brac, tramtrack, broad-complex zinc finger) proteins in the human genome, representing approximately 3% of all C2H2-type zinc finger proteins and about a quarter of all BTB-domain-containing proteins.1, 2 Essentially all of the human BTB-ZF proteins contain a single N-terminal BTB domain and a cluster of ZF motifs near the C-terminus, linked by a middle region that has low sequence conservation and is predicted to be unstructured. BTB-ZF proteins are often involved in developmental processes in a variety of cell types, and their expression is often regulated in a temporal and cell-type-specific fashion.3, 4, 5 For example, the promyelocytic leukemia zinc finger (PLZF) and Fanconi anemia zinc finger (FAZF) proteins are involved in regulating the differentiation of various cells in the hematopoietic lineage and in the testis.5, 6, 7, 8 Many BTB-ZF proteins have been characterized as sequence-specific transcription repressors, and some BTB-ZF proteins such as Miz1 (Myc-interacting zinc finger 1) act as either transcription repressors or activators, depending on their interaction partners.9, 10, 11, 12 Genetic lesions leading to abnormal fusion proteins or misregulation of the expression of BTB-ZF transcription factors have been observed in a variety of human cancers. Well-characterized examples include Bcl6 and Zbtb7/Lrf in B-cell lymphoma, and Zbtb16/Plzf in acute promyelocytic leukemia.

The BTB domain has at least two distinct roles in BTB-ZF proteins: the domain drives the dimerization of the protein via BTB–BTB homomeric associations13, 14 and couples the proteins to other factors via heteromeric protein–protein interactions.15, 16 The homodimerization of BTB domains appears to be a common property, while heterodimerization and oligomerization may occur with specific BTB domains. For example, Miz1 and BCL6 interact in the regulation of p21CIP1 and Bcl2 promoters, possibly through BTB–BTB interactions.10, 17

The BTB domains from several BTB-ZF proteins can function as autonomous trans-acting transcriptional repression modules; these activities are generally due to interactions with non-BTB proteins. Despite similar activities of many BTB domains from transcription factors, the landscape of BTB–corepressor interactions is complex. For example, the BTB domains from BCL6, PLZF, and Kaiso are able to recruit chromatin remodeling complexes via interactions with SMRT, N-CoR, BCOR, Sin3A, CtBP, and/or histone deacetylases.15, 18, 19, 20, 21 BCL6BTB binds to BBD motifs in N-CoR, SMRT, and BCOR,15, 16 while LRFBTB does not interact with SMRTBBD.14 KaisoBTB and Hic-1BTB do not interact with SMRT,21, 22 but KaisoBTB interacts with N-CoR21 and Hic-1BTB interacts with CtBP.23 The FAZF protein, encoded by the Zbtb32 gene (also known as TZFP,24 PLZP,7 and ROG25, 26), is a transcriptional repressor24, 26 and close sequence neighbor of PLZF.27, 28 However, it is not clear whether repression is dependent on an SMRT/N-CoR–histone deacetylase complex.18, 24, 25 Miz1 interacts with c-Myc via residues in the ZF region of the protein29 and is a component of both activation and repression complexes.11, 12

The structures of several BTB domains are known. The BTB domains from PLZF, BCL6, and LRF are tightly interwound homodimers.13, 14, 15, 30, 31 Each structure consists of a core BTB fold preceded by a 25-amino-acid to 30-amino-acid N-terminal extension that interacts mainly with the partner chain in the dimer, suggestive of domain swapping.32 This region forms a β-strand and an α-helix that make up a large part of the interchain interface.2 Truncation of a portion of the N-terminus of PLZFBTB resulted in misfolded protein33; however, in BACH1 and BACH2 BTB domains, residues N-terminal to the first β-strand may be important for regulating BTB self-association.34

In this study, we determined the crystal structures of the BTB domains from FAZF and Miz1 to 2.0 Å and 2.6 Å, respectively. These two BTB domains provide insight into the structural role of the N-terminal elements of the BTB domain. Miz1 was of particular interest because it lacks the residues required for forming the N-terminal β-strand, and we found an unexpected variability in the conformation of the amino-terminal region of FAZF. We analyze the dimer interfaces and lateral groove region of these BTB domain dimers in light of previous structures. Finally, we construct a molecular phylogeny of the BTB fold, providing evidence that the modified N-terminal regions of FAZF and Miz1 are more recent developments in the evolution of the domain.

Section snippets

FAZFBTB is not domain swapped

The asymmetric unit of the FAZFBTB crystal contains two crystallographically independent chains, each forming symmetric homodimers through crystallographic 2-fold axes (Fig. 1a and b, Table 1). The two chains have very similar structures, with an RMSD of 0.65 Å among all Cα atoms. As seen with other BTB domains, FAZFBTB is a butterfly-shaped homodimer that is predominantly α-helical, with flanking β-sheets at the “top” and “bottom” of the dimer (Fig. 1b). The two crystallographically

Conclusions

In this study, we report further insights into the structure and function of the BTB domain with the crystal structures of FAZFBTB and Miz1BTB. These structures both reveal the homodimer BTB fold, but each contains conformational changes. FAZFBTB is the first demonstration of a nonswapped BTB homodimer. Miz1BTB resembles a domain-swapped dimer even though it lacks the β1 element and also shows a displacement of 10 residues corresponding to β4. Even with these movements, the FAZF and Miz1 BTB

Protein expression and purification

cDNA encoding full-length human Miz1 and FAZF were obtained as Mammalian-Gene-Collection-compliant IMAGE clones (Mammalian Gene Collection IDs 161441 and 21109, respectively) from ATCC. The regions coding for the BTB domain (residues 1–115 of Miz1 and residues 1–117 of FAZF) were subcloned into a modified pET-32(a) vector coding for a thioredoxin His6-tagged fusion protein. BTB domains were purified and stored as previously described,13, 14 with an identical buffer [300 mM NaCl, 20 mM Tris

Acknowledgements

We thank Alexa Bramall for the initial purification and crystallization trials of FAZFBTB, Sylvia Ho for performance of analytical ultracentrifugation experiments and model fitting, Kosta Popovic for help with crystal handling, and the beamline staff at Cornell High Energy Synchrotron Source F2 and Advanced Photon Source 19-ID. We also thank Philip Bourne and Eric Scheeff for helpful comments on combining sequence and structure information for the construction of phylogenetic trees. This

References (57)

  • LinW. et al.

    Identification and gene structure of a novel human PLZF-related transcription factor gene, TZFP

    Biochem. Biophys. Res. Commun.

    (1999)
  • HoatlinM.E. et al.

    A novel BTB/POZ transcriptional repressor protein interacts with the Fanconi anemia group C protein and PLZF

    Blood

    (1999)
  • SchubotF.D. et al.

    Structure of the POZ domain of human LRF, a master regulator of oncogenesis

    Biochem. Biophys. Res. Commun.

    (2006)
  • GronenbornA.M.

    Protein acrobatics in pairs—dimerization via domain swapping

    Curr. Opin. Struct. Biol.

    (2009)
  • SteadM.A. et al.

    A beta-sheet interaction interface directs the tetramerisation of the Miz-1 POZ domain

    J. Mol. Biol.

    (2007)
  • KeskinO. et al.

    Hot regions in protein–protein interactions: the organization and contribution of structurally conserved hot spot residues

    J. Mol. Biol.

    (2005)
  • BoganA.A. et al.

    Anatomy of hot spots in protein interfaces

    J. Mol. Biol.

    (1998)
  • LiX. et al.

    Overexpression, purification, characterization, and crystallization of the BTB/POZ domain from the PLZF oncoprotein

    J. Biol. Chem.

    (1997)
  • JohnsonM.L. et al.

    Analysis of data from the analytical ultracentrifuge by nonlinear least-squares techniques

    Biophys. J.

    (1981)
  • PetreyD. et al.

    GRASP2: visualization, surface properties, and electrostatics of macromolecular structures and sequences

    Methods Enzymol.

    (2003)
  • BatemanA. et al.

    The Pfam protein families database

    Nucleic Acids Res.

    (2004)
  • StogiosP.J. et al.

    Sequence and structural analysis of BTB domain proteins

    Genome Biol.

    (2005)
  • CostoyaJ.A.

    Functional analysis of the role of POK transcriptional repressors

    Brief Funct. Genomics Proteomics

    (2007)
  • PiazzaF. et al.

    Disruption of PLZP in mice leads to increased T-lymphocyte proliferation, cytokine production, and altered hematopoietic stem cell homeostasis

    Mol. Cell. Biol.

    (2004)
  • DoulatovS. et al.

    PLZF is a regulator of homeostatic and cytokine-induced myeloid development

    Genes Dev.

    (2009)
  • PhanR.T. et al.

    BCL6 interacts with the transcription factor Miz-1 to suppress the cyclin-dependent kinase inhibitor p21 and cell cycle arrest in germinal center B cells

    Nat. Immunol.

    (2005)
  • ChanC.H. et al.

    Deciphering the transcriptional complex critical for RhoA gene expression and cancer metastasis

    Nat. Cell Biol.

    (2010)
  • HerkertB. et al.

    The Arf tumor suppressor protein inhibits Miz1 to suppress cell adhesion and induce apoptosis

    J. Cell Biol.

    (2010)
  • Cited by (21)

    • Identification of an atypical interaction site in the BTB domain of the MYC-interacting zinc-finger protein 1

      2021, Structure
      Citation Excerpt :

      Only one out of six BTB subunits extracted from these structures contains an inherent B3 strand (Figure S2); in this case, however, the strand is part of an intermolecular β sheet within a crystallographic tetramer, which we and others did not detect in solution (Stead et al., 2007; Stogios et al., 2010). In sum, these analyses corroborate the notion that the lack of a pre-formed B3 strand and associated plasticity of the B3-region in MIZ1BTB do not result from crystal packing but represent a specific property of this BTB domain (Stogios et al., 2010). In line with the crystal structures we determined and previous studies (Stogios et al., 2010), SEC-coupled multi-angle light scattering (MALS) analyses showed that MIZ1BTB is dimeric in solution (Figure S3A).

    • Adaptor protein self-assembly drives the control of a cullin-RING ubiquitin ligase

      2012, Structure
      Citation Excerpt :

      This behavior was not observed for SPOPBTB, which maintained a dimeric state across all examined concentrations (Figure 4E). The absence of monomeric species could be accounted for by the very low dimer dissociation constant that has been observed for other BTB domains (Stogios et al., 2010). Taken together, these data indicate that the BACK domain can promote formation of large (>500 kDa), multimeric (>25mers), dynamic, and dissociable SPOP assemblies in solution.

    • Drosophila BTB/POZ domains of "ttk Group" can form multimers and selectively interact with each other

      2011, Journal of Molecular Biology
      Citation Excerpt :

      Nondenaturing PAGE and cross-linking experiments were performed with 5 μM proteins (Figs. 3 and 4). Recently, a Kd of less than 2 μM for dimer formation was determined for three BTB domains;9 previously, a Kd of 110 μM was reported for dimer–tetramer association of Miz1.10 Therefore, our experiments show that multimers of the GAF BTB domain are stable in solution with a Kd of at least 5 μM; multimers of Mod(mdg4) and Tramtrack BTB domains appear to be even more stable; Batman forms less stable multimers, with Kd estimated to be between 10 and 60 μM.

    • Structural analysis of the PATZ1 BTB domain homodimer

      2020, Acta Crystallographica Section D: Structural Biology
    • Structures of heterodimeric POZ domains of Miz1/BCL6 and Miz1/NAC1

      2014, Acta Crystallographica Section F: Structural Biology Communications
    • Experimental phasing using zinc anomalous scattering

      2012, Acta Crystallographica Section D: Biological Crystallography
    View all citing articles on Scopus
    View full text