Crystal structure of the Bach1 BTB domain and its regulation of

Crystal structure of the Bach1 BTB domain and its regulation of homodimerization Blackwell X-ray N Ito et Structure al.Publishing of Bach1 Inc BTB domain homodimer

Nobutoshi Ito1,3a, Miki Watanabe-Matsui2, Kazuhiko Igarashi2 and Kazutaka Murayama1,3*b 1

Biomedical Engineering Research Organization, Tohoku University, 2-1 Seiryo-machi, Aoba-ku, Sendai 980-8575, Japan Department of Biochemistry, Graduate School of Medicine, Tohoku University, 2-1 Seiryo-machi, Aoba-ku, Sendai 980-8575, Japan 3 Protein Research Group, Genomic Sciences Center, RIKEN Yokohama Institute, 1-7-22 Suehiro-cho, Tsurumi, Yokohama 230-0045, Japan 2

The BTB/POZ domain is known as a protein–protein interaction motif that mediates homodimer and higher order self-associations. Proteins containing the BTB domain exist throughout eukaryotes; however, there is little information about the mechanism that determines the oligomeric state of the BTB domain. To address this question, we have determined the X-ray structure of the mouse Bach1 BTB domain. The present structure is similar to the previously determined BTB domain folds, including the human Bach1 BTB domain; however, distinct structural features are present, such as a novel homodimer interaction surface. The homodimer formation was found to involve a novel hydrogen bond network and interactions between hydrophobic surfaces of the kinked N-terminus (N-hook) and the partner’s C-terminal residues. The deletion of the N-hook resulted in the conversion of the homodimer into a monomer in solution, indicating that the N-hook promotes the homodimerization of the mBach1 BTB domain. We have also found that the BTB domain of Bach2, a protein highly related to Bach1, is present as a monomer due to a short peptide insertion at the N-hook. These results represent the first example of the key modulatory element of BTB domain homodimerization.

Introduction The BTB (bric-a-brac, tramtrack, and broad complex) domain, also known as the POZ (Pox virus and zinc finger) domain, is found in more than 500 proteins throughout eukaryotes. It is a protein–protein interaction motif that mediates homodimer and higher order self-associations as well as interactions with a number of other proteins (Stogios et al. 2005; Perez-Torrado et al. 2006). BTBcontaining proteins are reportedly involved in transcriptional repression (Melnick et al. 2000; Ahmad et al. 2003), cytoskeleton regulation (Bomont et al. 2000; Ziegelbauer et al. 2001; Kang et al. 2004), ion channel gating Communicated by: Shunsuke Ishii *Correspondence: [email protected] a Present address: Nagasaki University School of Medicine, 1-12-4 Sakamoto, Nagasaki 852-8523, Japan. b Present address: Graduate School of Biomedical Engineering, Tohoku University, 2-1 Seiryo-machi, Aoba-ku, Sendai 980-8575, Japan.

by tetramerization (Kreusch et al. 1998; Minor et al. 2000) and protein ubiquitination for degradation (Furukawa et al. 2003; Geyer et al. 2003; Krek 2003; Pintard et al. 2003; Xu et al. 2003; Kobayashi et al. 2004; Pintard et al. 2004; Wilkins et al. 2004; Willems et al. 2004). Depending on the spatial organization of secondary structure elements and the types of protein–protein associations, proteins with the BTB domain are classified into four families, including the BTB-zinc finger (ZF), Skp1, Elongin C and voltage-gated potassium channel T1 (T1-Kv) proteins (Stogios et al. 2005; Perez-Torrado et al. 2006). The BTB-ZF family contains BTB domain with an amino-terminal extension and forms homodimers (Ahmad et al. 1998, 2003), while the Skp1 proteins contain a family-specific carboxy-terminal extension to BTB domain and exist as monomers in heterotrimeric SCF (Skp1-Cullin-F-box-like E3 ubiquitin ligase) complexes (Zheng et al. 2002; Orlicky et al. 2003; Wu et al. 2003). The Elongin C and T1-Kv proteins consist of only the core BTB domain fold. Elongin C exists as a monomer in the heterotrimeric VHL-Elongin C-Elongin

DOI: 10.1111/j.1365-2443.2008.01259.x © 2009 The Authors Journal compilation © 2009 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

Genes to Cells (2009) 14, 167–178

167

N Ito et al.

B complex, whereas the T1 domains in T1-Kv proteins form homotetramers (Kreusch et al. 1998; Stebbins et al. 1999; Botuyan et al. 2001; Nanao et al. 2003). While the tertiary structures of the BTB domains are quite similar, the amino acid sequence similarity between members of different families is low. Therefore, the BTB domain fold is a versatile scaffold that participates in a variety of family-specific protein–protein interactions. The crystal structures of the BTB domains of B-cell lymphoma 6 (BCL6) (Ahmad et al. 2003; Ghetu et al. 2008), promyelocytic leukemia zinc finger protein (PLZF) (Ahmad et al. 1998; Li et al. 1999), leukemia/lymphoma related factor (LRF; also known as OCZF, FBI-1, and pokemon) (Schubot et al. 2006; Stogios et al. 2007), and Myc-interacting zinc-finger protein (Miz-1) (Stead et al. 2007) were previously reported. These studies provided a structural framework for the BTB domain. These BTB domain structures form homodimers by strand-exchanged domain-swapping, which is accomplished by the association of the N- and the partner’s C-terminal strands, to form a two-stranded antiparallel sheet. It was proposed that 3D domain swapping could be a general mechanism for switching to an oligomer from a monomer (Liu & Eisenberg 2002). However, there is no evidence for monomer–dimer exchange by the BTB domain. It is also not known how the specificity of the BTB–BTB interaction is achieved. Bach1, the subject of this article, is a transcriptional repressor of heme-related genes, such as the globin and heme oxygenase-1 genes (Igarashi & Sun 2006). Bach1 and its related factor, Bach2, are unique among the BTB domain proteins in that they contain a CNC (cap’n’collar) type basic leucine zipper DNA binding domain (bZip) (Oyake et al. 1996). Bach1 binds to a Maf recognition element (MARE) by forming heterodimers with small Maf proteins, such as MafK. The Bach1/ MafK heterodimers interact with each other through the Bach1 BTB domain, forming a multivalent DNA binding complex (Igarashi et al. 1998; Yoshida et al. 1999). Bach1 and Bach2 are found only in vertebrates, whereas their ancestral gene, with the BTB-bZip configuration, is present in chordates (Amoutzias et al. 2007). These results imply that a BTB-bZip type transcriptional repressor has a novel structural feature to define the molecular function. In this study, we determined the X-ray structure of the mouse Bach1 BTB domain (mBach1 BTB), constructed from the first methionine, to elucidate the structural basis of the BTB-bZip protein. The present structure revealed novel BTB domain homodimer interactions, which were not present in the previously determined crystal structure of human Bach1 BTB domain (hBach1 BTB) deposited in the Protein Data Bank (PDB ID 2IHC). Based on 168

Genes to Cells (2009) 14, 167–178

structural and biochemical analyses, we elucidated how the oligomeric state is maintained in the Bach family BTB domains, including Bach2. Our results provide novel insight into the structural determinants for the dimerization of the Bach family BTB domain.

Results Overall structure of mBach1 BTB domain

We have crystallized and solved the structure of the mBach1 BTB (residues 1–135, one molecule per asymmetric unit) at 2.5 Å resolution (Fig. 1A). To obtain the initial phases by molecular replacement, the human homolog (residues 7–126) was used as the search model. Although the model structure lacks the N- and C-terminal residues, the electron density map calculated using the initial refinement model revealed the presence of additional densities consistent with the residues belonging to the N- and C-terminal regions, corresponding to residues 1–6 and 127–135 of mBach1 as well as the disordered loop (residues 65–69) in the hBach1 BTB. The mBach1 BTB consists of six helices (helices H1–H6) and five strands (strands S1–S5). As observed for all BTB domains, the mBach1 BTB structure reveals 3D-domain swapping. The N-terminal S1 region of each monomer interacts with its partner’s S5 strand, resulting in the formation of a strand-exchanged homodimer (Fig. 1A). Another interface is formed by the helices H1, H2, and H3 between the two subunits. According to the nomenclature of a domain-swapped protein, these interfaces are referred to as follows: the former is a “closed interface” and the latter is an “open interface” (Ahmad et al. 1998; Liu & Eisenberg 2002). The structural similarities of the mouse and human Bach1 BTB domains (one and four molecules in the asymmetric unit, respectively; sequence identity is 96.3%) are as follows: root mean square deviation values are 1.1 Å (chain A of 2ICH; 117 Cα atoms), 0.7 Å (chain B; 111 Cα atoms), 1.0 Å (chain C; 116 Cα atoms), and 1.1 Å (chain D; 115 Cα atoms). The root mean square deviation between the dimer of mouse and human is 0.88 Å, which indicates that the architecture of the dimer is basically identical to each other. In addition, a structure similarity search for other BTB domains using the DALI server revealed that the mBach1 BTB has the highest similarity to the PLZF protein, with a Z score of 16.4 and an root mean square deviation value of 1.5 Å (for 118 equivalent Cα atoms). The total contact surface of the homodimer is about 2150 Å2. It has been reported that the contact surface of other BTB domains are 1600– 2000 Å2 (Schubot et al. 2006).

© 2009 The Authors Journal compilation © 2009 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

X-ray structure of Bach1 BTB domain homodimer

Figure 1 Overall structure of the mouse Bach1 BTB domain. (A) Ribbon representation of the mBach1 BTB homodimer. The monomers are colored blue and green. The secondary structures are labeled. The figure was generated with Pymol (http://www.pymol.org). The open and closed dimerization interfaces are indicated by the solid and dashed boxes, respectively. (B) Structure-based sequence alignment of mouse Bach1 and Bach2, and the BTB proteins with determined structures, including PLZF, BCL6 and LRF. The secondary structures are indicated above (helices, rectangles; sheets, arrows; coil, solid lines). Identical and conserved residues are indicated as the red box and the red character, respectively. Mutated residues in this study are labeled with an underline and asterisk (*). The alignment was generated and illustrated using the programs ClustalW (Chenna et al. 2003) and ESPript (Gouet et al. 2003).

Hydrogen bond network at the open interfaces

After finding the homodimer formation, we next examined the effect of mutations at the open interfaces that form a buried surface. We prepared six alanine mutants, which were divided into two groups. The first group consists of Leu21, Gln27 and Asp35, which were studied

well in a previous report using the PLZF BTB domain (Melnick et al. 2000). In addition, Leu21 and Asp35 are invariant residues among the BTB domains with structures that have been determined (Fig. 1B). To distinguish each mutant, the previous nomenclature was used, such as “monomer core” for Leu21, “surface” for Gln27, and “charged pocket” for Asp35 (Melnick et al. 2000). Leu21


Genes to Cells (2009) 14, 167–178

169

N Ito et al.

Figure 2 Novel hydrogen bonds at the open interface. (A) Schematic representation of the mutated residues at the open interface. The side chains of the residues are represented by stick models. (B) Hydrogen bonds between Ser17 and Thr18, and Asn19 and Asn117# are shown by dotted lines (orange), and the distances are labeled. The molecule with a 180° horizontal rotation is shown as a reference.

Table 1 Summary of the experimental and theoretical molecular mass (Da) estimated by the ultracentrifugation analysis of the Bach1 and Bach2 BTB domains

Experimental Theoretical

Bach1

Bach1

Bach1

Bach1

Bach1

Bach2

wild-type 28 650 15 515

ΔDN6 14 317 15 025

S17A 20 541 15 174

T18A 17 835 15 160

N19A 16 542 15 147

wild-type 15 243 16 184

and Gln27 are located on helix H1, and Asp35 is present on the loop H1-S2 (Fig. 2A). When expressed in Escherichia coli, these Ala substituted mutants, Leu21Ala, Gln27Ala and Asp35Ala, formed inclusion bodies or aggregated during the purification, suggesting that these mutations destabilize the domain. Therefore, it was difficult to perform the biophysical analysis. These observations are consistent with the previous mutational analysis of the PLZF BTB domain (Melnick et al. 2000). The second group carried mutations at Ser17, Thr18 or Asn19. We chose these residues because we found that they were involved in a novel hydrogen bond network at the tip of the helix H1 (Fig. 2B). Symmetrical interactions provided by the side chains of Ser17-Thr18#, Ser17#Thr18, Asn19-Asn117# and Asn19#-Asn117 (“#” denotes the residues from the partner molecule) form the hydrogen bond networks. These reciprocal interactions connect the helices H1-H1#, and the helix H1 and the loop H5#–H6#, respectively (Fig. 2B), and seem to stabilize the homodimer. While the Asn19 and Asn117 residues are conserved between Bach1 and Bach2, the corresponding 170

Genes to Cells (2009) 14, 167–178

residues of other BTB domain structures are not conserved (Fig. 1B). Hydrogen bonding between helices H1–H1# is also found in the LRF BTB domain, whereas the Asn19–Asn117# interaction is unique to the Bach family. Therefore, we examined whether these residues were involved in stabilizing the dimerization, by examining the oligomeric state of their Ala mutants. In contrast to the former three mutants, the three mutants Ser17Ala, Thr18Ala and Asn19Ala were expressed as soluble proteins and were purified in a similar manner as the wild-type (WT) protein. Sedimentation equilibrium yielded an estimated molecular mass of 28 650 Da for the WT mBach1 BTB, which is close to twice the theoretical monomeric value of 15 514.5 Da (Fig. 3A), and 20 541, 17 835 and 16 542 Da for the Ser17Ala, Thr18Ala and Asn19Ala mutants, respectively (Fig. 3C– E and Table 1). These values are slightly larger than the theoretical molecular mass (15 515 Da) of the Bach1 BTB (Table 1). Therefore, the mutations in the open interfaces destabilized the dimer, resulting in a mixture of the monomer and homodimer. These results indicate



Figure 3 Analytical ultracentrifugation. A plot of the sedimentation equilibrium data with the residuals from the best fit to a single ideal species. The Bach1 WT, ΔN6, S17A, T18A, N19A BTB domains and Bach2 [1–138] are shown in boxes A–F, respectively. © 2009 The Authors Journal compilation © 2009 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

Genes to Cells (2009) 14, 167–178

171

N Ito et al.

The N-hook promotes homodimerization

Figure 4 Novel interactions at the closed interface. Hydrophobic interaction by the N-hook structure at the closed interface. The partner molecule is represented by a transparent molecular surface. Hydrogen bonds are represented by orange dotted lines.

that the newly identified hydrogen-bond network at the open interfaces also contributes to the homodimer formation. The closed interfaces show novel N-hook interactions

The present structure has the ordered short N- and Cterminal segments, allowing us to visualize several additional residues that contribute to the subunit-subunit interface. The N-terminal loop connected to sheet S1 (loop N-S1) associates with the loop protruding from helix H6 (loop H6-C) at the partner’s C-terminal region (Fig. 1A). In the C-terminal region of the partner molecule, the residues Phe125#, Phe128# and Phe130# form a hydrophobic surface (Fig. 4). The side chain of Phe9 is fitted into the surface upon binding, and as a result, the loop N-S1 is sharply kinked at Ser6 by about 90° (referred to as an N-hook). The N-hook interaction seems to be stabilized by the side chains of Val3 and Ala7, and the main chain backbone of Met1 forms hydrogen bonds with Lys127# and Phe128#. The structure suggests that the N- and C-terminal loops specifically associate with each other. In the previously reported studies of BTB domain structures, including PLZF, BCL6 and LRF, the equivalent regions for the N-hook were structurally disordered or artificially removed for the structural studies. It was proposed that the closed interface is mainly stabilized by the interaction between sheets S1 and S5 (Ahmad et al. 1998; Liu & Eisenberg 2002; Schubot et al. 2006; Stogios et al. 2007). These differences suggest that the N-hook of Bach1 has a role distinct from the closed interface of the canonical BTB domains. 172

Genes to Cells (2009) 14, 167–178

In order to examine whether the association of the Nhook is involved in the dimer formation in solution, we performed analytical ultracentrifugation analyses and estimated the oligomerization state. For the analyses, we prepared the WT and mutant lacking residues 2–6 of the mBach1 BTB domain (referred as BTBΔΝ6). As mentioned above, Sedimentation equilibrium for the WT mBach1 BTB is 28 650 Da. In contrast, the corresponding value for the mBach1 BTBΔN6 is 14 317 Da, which is close to the theoretical monomeric value of 15 025 Da (Fig. 3E and Table 1). These results indicated that the predominant species of the WT mBach1 BTB is the dimer, while that of the mBach1 BTBΔN6 is the monomer, although we cannot exclude the possibility that the remaining protease recognition site (GPMH) might negatively affect dimer formation. It is interesting that abolishment of interactions around the N-hook region can convert the mBach1 BTB to the monomeric form, because the BTB was considered to form a rigid homodimer (Melnick et al. 2000). In a previous study, the presence of a monomeric BTB domain in PLZF and BCL6 was not reported (Li et al. 1997). Furthermore, BTB monomers were considered to be intrinsically unstable, and mutations that destabilized the dimer interface caused misfolding (Melnick et al. 2000). Therefore, our result is the first observation that the BTB domain without the N-hook can exist as a stable monomer. This raises the possibility that a modification of the N-hook segment could regulate dimer formation and/or intersubunit exchange between different BTB molecules. However, at present, no modification of the N-terminus of the Bach1 BTB domain has been reported. Gel filtration analysis supports the different oligomerization modes provided by the N-hook interactions

We further investigated the oligomeric state of the protein by other biochemical methods. In the gel filtration chromatography, the WT protein appeared in the elution volume of 10.2 mL, which was earlier than Conalbumin (43 kDa) whose elution volume was 10.4 mL, and the elution peak was tailed to lower molecular weight range. The elution time of the WT Bach1 BTB domain was significantly faster than those of the Bach1 BTB DN6 (11.7 mL) and the alanine mutants (10.9, 10.7 and 10.8 mL for S17A, T18A and N19A, respectively). These data indicate that the mutants have prolonged retention times. The estimated molecular mass of the WT protein is about 52 kDa, which is approximately



threefold of the theoretical molecular mass and larger than the value obtained from the analytical ultracentrifugation. Furthermore, the molecular masses of every protein showed larger values than those obtained by the ultracentrifugation. Both the mass and shape of a protein affect the gel filtration positions. In the gel filtration, the molecular masses of the Bach1 BTB proteins might have appeared to be higher, due to the butterfly-like structural feature. Nonetheless, the oligomeric state obtained from gel filtration of each protein was quite consistent with ultracentrifugation. The relative values of the molecular masses of the WT, S17A, T18A and N19A mutants are 1.7-, 1.35-, 1.45- and 1.41-fold larger than that of the ΔDN6 mutant, respectively. Therefore, probably the oligomeric states among the WT, the ΔDN6 mutant, and the alanine mutants are different. These data are basically consistent with the above ultracentrifugation data and suggest that these mutations destabilized dimer formation. The N-hook is the key element for the oligomerization state of Bach1 and Bach2

We have shown that the N-hook truncated Bach1 BTB domain mutant can stably exist as the monomer (Table 1). This result suggests that the N-hook plays a key role to keep the oligomeric state of the Bach1 BTB domain. To investigate whether this element function in other BTB proteins, we compared the Bach1 and Bach2 BTB domains. Although the overall sequences of the Bach1 and Bach2 BTB domains are quite similar to each other (91% similarity and 59% identity for 138 residues), a significant difference exists at the N-terminal region of Bach2 that corresponds to the Bach1 N-hook. Mouse Bach2 has a short insertion (Lys6-Pro7-Gly8) in this region (position 8 is Asp in human) (Figs 1B and 5A). Interestingly, sedimentation equilibrium analyses yielded an estimated molecular mass of 15 243 Da for mBach2 BTB, which is close to the theoretical monomeric value of 16 184 Da at the physiological salt concentration (150 mm NaCl) (Fig. 3F and Table 1). We also examined the interaction between the glutathione S-transferase (GST)-fused and His-tagged Bach2 BTB domains by a pulldown analysis using glutathione Sepharose resin, under the same salt conditions as in the analytical centrifugation (Fig. 5B). The results confirmed that there was no significant homodimer formation by the Bach2 BTB domain. In the gel filtration, the relative value of the molecular mass of the Bach2 BTB domain was approximately 1.1-fold larger than that of the ΔN6 mutant. These data indicate that Bach2 BTB exists as a monomer in vitro. The different oligomeric states between the Bach1 and Bach2 BTB domains may be attributable

Figure 5 The Bach2 BTB domain is a monomer. (A) Sequence comparison of the N-hook segment of mouse (m) and human (h) Bach1 and Bach2. The significant inserted region is highlighted in the solid box. (B) GST-pulldown interaction analysis between the GST-Bach2 BTB and the His-tagged Bach2 BTB domain. GST (mock; lane2) or GST-Bach2 BTB (lane 3) was immobilized to GS4B resin, and was incubated with the His-tagged Bach2 BTB domain. After the beads were washed, the reactions were fractionated by SDS-PAGE, and the gel was stained with Coomassie Brilliant Blue. (Lanes 1: His-Bach2 [1–138] (10% input), 2: GST (mock), 3: GST-Bach2 [1–138], 4: Molecular Weight Markers).

to the N-hook sequence: the Bach2 BTB domain autonomously acts as a monomer because of the insertion, which perturbs the N-hook association. To investigate the biological function of the N-hookmediated dimerization, we examined the transcriptional repression activity using a luciferase reporter driven by MARE. When the reporter and expression plasmids were co-transfected and transiently over-expressed in NIH 3T3 cells, the expression of WT Bach1 and Bach1ΔN6 resulted in the repression of the luciferase activity in a dose-dependent manner (Fig. 6). This result indicates that Bach1ΔN6 can repress the expression of luciferase as


Genes to Cells (2009) 14, 167–178

173

N Ito et al.

Figure 6 Transcriptional repression of HO-1 enhancers by Bach1 and Bach1ΔN6. NIH 3T3 cells were transfected with the HO-1 reporter (0.5 μg), and with the Bach1 or Bach1ΔN6 expression plasmid (1, 3 or 9 ng).

well as the WT. Therefore, the deletion of the N-hook barely affected the transcriptional repression activity, suggesting that the mutant Bach1 with the monomeric BTB domain can also repress the gene expression. Indeed, it is assumed that the DNA binding activity of the bZip domain is principally responsible for the repression function: Bach1 can inhibit transcription by competing for MARE with MARE-binding transcription activators such as Nrf2 (Sun et al. 2004). Bach1ΔN6 has an intact bZip domain to form a complex with MafK, and the heterodimer is sufficient for the repression. Taken together, the dimerization of the BTB domain is not essential for the transcriptional repression in the reporter assays. However, the N-hook-mediated dimerization of the BTB domain may be either involved in the formation of a higher order transcription repressing complex, or required for modulating the dimer specificity by another factor. In fact, Bach2 has been purified as a multi-protein complex from cultured cells (Ochiai et al. 2006). The precise biological function of the N-hook should be explored in the future.

Discussion It is known that the electron densities of the both terminal regions cannot be identified frequently due to disorder, even if the crystallization constructs include these regions. The secondary structure prediction program (Bryson et al. 2005) indicated that the N-terminal regions (before the first β-sheet, S1) of both Bach proteins are also flexible. Although this region has not been paid special attention in BTB domain structures so far, the crystal structure of the mBach1 BTB demonstrated that the N-hook plays an important role for multimerization with the novel 174

Genes to Cells (2009) 14, 167–178

interactions in the vicinity of the closed interface of the domain-swapped dimer. The specific interaction made by the N-hook suggests the importance of maintaining rigidity in the dimerized BTB domain. For instance, the removal of the N-hook converts the Bach1 BTB domain homodimer into a monomer in solution. The importance of the N-hook was further supported by the finding that the Bach2 BTB domain, which has the insertion in the N-hook, existed as a monomer in solution. These structural and biophysical studies revealed that the N-hook plays a pivotal role in the stabilization of the dimer and functions as one of structural key elements for homodimer formation by the Bach1 BTB domain. Our results also indicate that the Bach family BTB domains are interesting examples for understanding the principles of the oligomerization and stabilization of the numerous BTB domains, with their various multimerization modes. Our biochemical data revealed that the Bach1 and Bach2 BTB domains showed different oligomeric properties in vitro, and may be related to the previous finding that each BTB domain displays different activities in living cells (Tashiro et al. 2004). In addition, the BTB domain of Bach1 mediates long range interaction among cis-DNA elements, forming DNA loops observed in the atomic force microscopy (Yoshida et al. 1999). The atomic force microscopy study estimated that eight Bach1-MafK heterodimers (total 16 molecules) associate to form the DNA loop (Yoshida et al. 1999). To assemble the DNAloop structure, it is likely that both homomulitimerization through Bach1 BTB domain and the specific DNA binding by bZip domains of Bach1-MafK heterodimer (or other small Maf proteins) are required. Therefore, the stable BTB homodimer as well as the bZip interaction will be important and minimal unit to mediate the long range interaction among respective cis-DNA elements in the genome. The N-hook assisted dimerization of the BTB domains is expected to contribute to the construction of this minimal unit. At present, however, we cannot generalize which BTB domains are governed by the N-hook mechanism, as there are few examples available for comparisons of the sequence features and the tertiary structures of the Nhook. Therefore, identifying functional N-hooks in other BTB domains and clarifying the sequence motifs will be the subjects for future research.

Experimental procedures Plasmid construction For the structural and biochemical assays, the cDNA clones encoding mouse Bach1 (amino acid residues 1–135 and 7–135)


X-ray structure of Bach1 BTB domain homodimer and Bach2 (residues 1–138) were amplified by PCR with primers that included NdeI and BamHI sites. The respective cDNAs were subcloned into modified pET28a (Novagen) and pGEX6p-1 (GE Healthcare) plasmids, containing the rhinovirus 3C protease recognition site and NdeI and BamHI cloning sites. Ala scanning mutants were generated using the Altered Sites II in vitro Mutagenesis System (Promega), according to the manufacturer’s protocol. The procedures used for constructing the mutant proteins in the plasmids were the same as those reported for the WT mBach1 BTB. For the luciferase assay, mammalian expression plasmids for the full-length and ΔN6 mouse Bach1 were constructed by isolating the corresponding ORF, using PCR with primers that included BclI and HindIII sites, and inserting the fragment into the BamHI and HindIII sites of pcDNA3.1FLAG (Muto et al. 2002). All constructs were verified by DNA sequencing.

Protein expression and purification The expression plasmids were transformed into E. coli BL21 (DE3). Cells were grown in LB media with 50 μg/mL kanamycin sulfate at 37 °C. Expression of the recombinant protein was initiated at A600 = 0.7–1.0, by adding isopropyl-β-d-thiogalactopyranoside (IPTG) to a final concentration of 1 mm and continuing the culture for 6 h at 30 °C. Cells were harvested by centrifugation and were stored at −80 °C. Frozen cells were suspended in 25 mm Hepes–NaOH (pH 8), 1 m NaCl, 10% glycerol, 1 mm dithiothreitol (DTT), 1 mm Tris (2-carboxyethyl) phosphine hydrochloride (TCEP) and 0.05% Nonidet P40 (NP40), and then were disrupted by sonication. The lysate was centrifuged to remove the cell debris. The soluble protein was purified with a HisTrap column (GE Healthcare) with an imidazole gradient, and the N-terminal hexa-Histidine tag was removed by PreScission ProteaseTM (GE Healthcare), leaving three nonnative amino acids, Gly-Pro-His, at the N terminus. The protein was further purified using anion exchange and Superdex75 chromatography (GE Healthcare), and the peak fractions eluted from the column were concentrated with an Amicon Ultra filter (Millipore).

Crystallization The purified protein in buffer (25 mm Hepes–NaOH (pH 8), 0.3 m NaCl, 1 mm DTT and 1 mm TCEP), concentrated to 7 mg/mL, was used for the initial crystallization screening by the sitting drop vapor diffusion method at 20 °C. The protein solution was mixed with an equivalent volume of mother liquor from several screening conditions, and was equilibrated against 0.1 mL of the mother liquor. The initial crystal was obtained under the following conditions (1 m (NH4)2HPO4 and 0.1 m imidazole (pH 8)) using a WizardTM kit (Emerald BioSystems). The best crystal was obtained by the micro seeding method, using a Seed Bead Kit (Hampton Research). The crystals typically grew to full size within one month. The crystal belonged to the space group P6122, with unit cell dimensions a = b = 106.2 Å and c = 91.0 Å.

Table 2 Crystallographic data and refinement statistics Diffraction data X-ray source Wavelength (Å) Resolution (Å) Unique reflections Redundancy Completeness (%)* I/σ(I)* Rsym (%)*† Molecular replacement statistics R-factor (%) Correlation function Refinement statistics Resolution (Å) Rwork/Rfree (%)‡ Protein atoms Water atoms Overall B-factor (Å2) Rmsd. Bond length (Å) Bond angles (°) Ramachandran (%) Most favored and additional allowed Generously allowed Disallowed

Photon factory/BL17A 1.0 50-2.5 11 031 10.3 99.7 (99.5) 13.4 (4.3) 7.0 (53.8) 58 0.443 20.0–2.5 22.9/26.2 1076 8 50.0 0.009 1.5 98.5 1.5 0

*Numbers in parentheses correspond to the values in the highest resolution shell. † Rsym = Σh Σi|Ihi – |/Σh Σi|Ihi|, where h are unique reflection indices, and i indicates symmetry equivalent indices. ‡ Rwork = Σ|Fo – Fc|/Σ Fo for overall data used in the refinement and Rfree = Σ|Fo – Fc|/Σ Fo, calculated with 5% of data excluded from refinement.

Data collection, structure determination and refinement Crystals were harvested by the addition of the mother liquor to crystal drops, and then were cryo-protected by sequential transfer into mother liquor containing 10% increments of ethylene glycol, to a final concentration of 20%. The crystals were flash-cooled and stored in liquid nitrogen. Diffraction data from the native crystals were collected to 2.5 Å at the Photon Factory beamline BL17A (Tsukuba, Japan). X-ray diffraction data were collected at 100 K, using a nitrogen stream, with a Quantum 4R CCD detector (Area Detector Systems Corporation). Diffraction data were processed with the programs DENZO and SCALEPACK in the HKL2000 program suite (Otwinowski & Minor 1997) (Table 2). The structure was determined by the molecular replacement method, using the program Molrep with the human Bach1 BTB (PDB ID; 2IHC) domain as the search model. The electron density


Genes to Cells (2009) 14, 167–178

175

N Ito et al. map was improved by density modification using the program DM (Cowtan 1994), by applying an estimated solvent content of 70%. Models were refined through alternating cycles using the programs O ( Jones et al. 1991) and CNS (Brunger et al. 1998). At an early stage of the refinement, the disulfide linkage between Cys107 and Cys122 was not included in the model. In the final refinement step, the disulfide bond was included in the model, and the Rwork/free-value was further improved. In the current model, the crystallographic Rwork/free converged to 22.9/26.2% (Table 2). The current model contains one molecule of the mBach1 BTB and eight water molecules. The Ramachandran plot of the final structural model showed 98.5% of the residues in the most favorable and additionally allowed regions and 1.5% of the residues in the generously allowed regions (Laskowski et al. 1993). The atomic coordinates and structure factors have been deposited in the Protein Data Bank (accession code 2Z8H).

Analytical ultracentrifugation The protein samples were purified as above, and the buffer was exchanged by gel filtration chromatography to buffer A (25 mm Hepes–NaOH (pH 8.0), 0.15 m NaCl, 1 mm TCEP, and 10% glycerol). Sedimentation equilibrium experiments were performed in a Beckman Optima XL-I instrument with six-channel centerpieces, using a Beckman An-50Ti rotor. The protein concentrations of the mBach1 BTB and mBach1ΔΝ6 proteins loaded into the cells were 3.2 and 3.1 mg/mL, respectively. Equilibrium distributions were analyzed after 17 h of centrifugation at 39 030 g rpm and at 4 °C. For the molecular weight analysis, partial specific volume and solution density values of 0.73 (0.723, only for Bach2 BTB domain) and 1.05 g/cm3, respectively, were used.

Luciferase reporter assay NIH 3T3 cells were grown in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum, 1% penicillin and 1% streptomycin, and were transfected with the HO-1 reporter plasmid (0.5 μg) (Alam et al. 2000) and with the Bach1 or Bach1ΔN6 expression plasmid (1, 3 or 9 ng) using the GeneJuice Transfection Reagent (Merck), according to the manufacturer’s instructions. Each transfection was done in duplicate, and the luciferase activity was measured 24 h after transfection, using the dual-luciferase reporter assay kit (Promega) with a Biolumat luminometer (Berthold), according to the manufacturers’ protocol. The fire-fly luciferase activity was normalized for transfection efficiency, as determined by the control sea pansy luciferase activity. The normalized values are reported as the mean ± SD from two independent experiments.

GST pull-down analysis Escherichia coli BL21 (DE3) cells were transformed with the GSTfusion Bach2 [1–138] protein expression plasmid and the GST parent vector, as a control. The cells were grown in LB medium supplemented with 100 μg/ml ampicillin at 37 °C. The expression of recombinant proteins was induced at A600 = 0.7–1.0 with 0.5 mm IPTG for 6 h at 30 °C. For His-Bach2 [1–138], the protein

176

Genes to Cells (2009) 14, 167–178

samples were purified as above, and the buffer was exchanged by gel filtration chromatography to buffer A. The GST-Bach2 [1–138] and GST (mock) proteins were purified on glutathione Sepharose 4B affinity beads (20 μL) that had been extensively washed with buffer A and were then incubated with the purified His-Bach2 [1–138] (10 μg) for 60 min at 4 °C. The beads were recovered by centrifugation at 13 000 rpm for 3 min and were washed four times with 1 mL of buffer A. An aliquot of the eluted proteins was resolved by SDS-PAGE and visualized by Coomassie Brilliant Blue staining.

Acknowledgements The authors thank Ryogo Akasaka for performing the analytical ultracentrifugation and Akiko Kanatsuka for technical assistance. Authors also thank the beamline staffs at BL17A and NW12 of the Photon Factory (Tsukuba, Japan) for technical help during data collection. This work was performed with the approval of the Photon Factory Program Advisory Committee (Proposal No. 2006G383). This work was supported by grants from Special Coordination funds for Promoting Science and Technology from the Japan Science Technology Agency. Part of this study was supported by grants-in-aid from the ministry of Education, Culture, Sport, Science and Technology of Japan, and Exploratory Research Program for Young Scientists of Tohoku University.

References Ahmad, K.F., Engel, C.K. & Prive, G.G. (1998) Crystal structure of the BTB domain from PLZF. Proc. Natl. Acad. Sci. USA 95, 12123–12128. Ahmad, K.F., Melnick, A., Bouchard, D., Liu, J., Kiang, C.L., Mayer, S., Takahashi, S., Licht, J.D. & Prive, G.G. (2003) Mechanism of SMRT corepressor recruitment by the BCL6 BTB domain. Mol. Cell 12, 1551–1564. Alam, J., Wicks, C., Stewart, D., Gong, P., Touchard, C., Otterbein, S., Choi, A.M., Burow, M.E. & Tou, J. (2000) Mechanism of heme oxygenase-1 gene activation by cadmium in MCF-7 mammary epithelial cells. Role of p38 kinase and Nrf2 transcription factor. J. Biol. Chem. 275, 27694–27702. Amoutzias, G.D., Veron, A.S., Weiner, J. III, Robinson-Rechavi, M., Bornberg-Bauer, E., Oliver, S.G. & Robertson, D.L. (2007) One billion years of bZIP transcription factor evolution: conservation and change in dimerization and DNA-binding site specificity. Mol. Biol. Evol. 24, 827–835. Bomont, P., Cavalier, L., Blondeau, F., Ben Hamida, C., Belal, S., Tazir, M., Demir, E.,Topaloglu, H., Korinthenberg, R., Tuysuz, B., Landrieu, P., Hentati, F. & Koenig, M. (2000) The gene encoding gigaxonin, a new member of the cytoskeletal BTB/kelch repeat family, is mutated in giant axonal neuropathy. Nat. Genet. 26, 370–374. Botuyan, M.V., Mer, G., Yi, G.S., Koth, C.M., Case, D.A., Edwards, A.M., Chazin, W.J. & Arrowsmith, C.H. (2001) Solution structure and dynamics of yeast elongin C in complex with a von Hippel-Lindau peptide. J. Mol. Biol. 312, 177–186. Brunger, A,T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P.,


X-ray structure of Bach1 BTB domain homodimer Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T. & Warren, G.L. (1998) Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D 54, 905–921. Bryson, K., McGuffin, L.J., Marsden, R.L., Ward, J.J., Sodhi, J.S. & Jones, D.T. (2005) Protein structure prediction servers at University College London. Nucl. Acids Res. 33, W36–W38. Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T.J., Higgins, D.G. & Thompson, J.D. (2003) Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res. 31, 3497–3500. Cowtan, K. (1994) “dm”: An automated procedure for phase improvement by density modification. Joint CCP4 and ESFEACBM Newsletter on Protein Crystallography 31, 34–38. Furukawa, M., He, Y.J., Borchers, C. & Xiong, Y. (2003) Targeting of protein ubiquitination by BTB-Cullin 3-Roc1 ubiquitin ligases. Nat. Cell Biol. 5, 1001–1007. Geyer, R., Wee, S., Anderson, S., Yates, J. & Wolf, D.A. (2003) BTB/POZ domain proteins are putative substrate adaptors for cullin 3 ubiquitin ligases. Mol. Cell 12, 783–790. Ghetu, A.F., Corcoran, C.M., Cerchietti, L., Bardwell, V.J., Melnick, A. & Privé, GG. (2008) Structure of a BCOR corepressor peptide in complex with the BCL6 BTB domain dimer. Mol. Cell 29, 384–391. Gouet, P., Robert, X. & Courcelle, E. (2003) ESPript/ENDscript: extracting and rendering sequence and 3D information from atomic structures of proteins. Nucleic Acids Res. 31, 3320– 3323. Igarashi, K., Hoshino, H., Muto, A., Suwabe, N., Nishikawa, S., Nakauchi, H. & Yamamoto, M. (1998) Multivalent DNA binding complex generated by small Maf and Bach1 as a possible biochemical basis for b-globin locus control region complex. J. Biol. Chem. 273, 11783–11790. Igarashi, K. & Sun, J. (2006) The heme-Bach1 pathway in the regulation of oxidative stress response and erythroid differentiation. Antioxid. Redox Signal. 8, 107–118. Jones, T.A., Zou, J.Y., Cowan, S. & Kjeldgaard, M. (1991) Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47, 110–119. Kang, M.I., Kobayashi, A., Wakabayashi, N., Kim, S.G. & Yamamoto, M. (2004) Scaffolding of Keap1 to the actin cytoskeleton controls the function of Nrf2 as key regulator of cytoprotective phase 2 genes. Proc. Natl. Acad. Sci. USA 101, 2046–2051. Kobayashi, A., Kang, M.I., Okawa, H., Ohtsuji, M., Zenke, Y., Chiba, T., Igarashi, K. & Yamamoto, M. (2004) Oxidative stress sensor Keap1 functions as an adaptor for Cul3-based E3 ligase to regulate proteasomal degradation of Nrf2. Mol. Cell. Biol. 24, 7130–7139. Krek, W. (2003) BTB proteins as henchmen of Cul3-based ubiquitin ligases. Nat. Cell Biol. 5, 950–951. Kreusch, A., Pfaffinger, P.J., Stevens, C.F. & Choe, S. (1998) Crystal structure of the tetramerization domain of the Shaker potassium channel. Nature 392, 945–948. Laskowski, R.A., McArthur, M., Moss, D.S. & Thoronton, J.M.

(1993) PROCHECK-a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26, 283–291. Li, X., Lopez-Guisa, J.M., Ninan, N., Weiner, E.J., Rauscher, F.J. III & Marmorstein, R. (1997) Overexpression, purification, characterization, and crystallization of the BTB/POZ domain from the PLZF oncoprotein. J. Biol. Chem. 272, 27324–27329. Li, X., Peng, H., Scultz, D.C., Lopez-Guisa, J.M., Rauscher, F.J. & Marmorstein, R. (1999) Structure-function studies of the BTB/POZ transcriptional repression domainfrom the promyelocytic leukemia zinc finger onco protein. Cancer Res. 59, 5275–5282. Liu, Y. & Eisenberg, D. (2002) 3D domain swapping: as domains continue to swap. Protein Sci. 11, 1285–1299. Melnick, A., Ahmad, K.F., Arai, S., Polinger, A., Ball, H., Borden, K.L., Carlile, G.W., Prive, G.G. & Licht, J.D. (2000) In-depth mutational analysis of the promyelocytic leukemia zinc finger BTB/POZ domain reveals motifs and residues required for biological and transcriptional functions. Mol. Cell. Biol. 20, 6550–6567. Minor, D.L., Lin, Y.F., Mobley, B.C., Avelar, A., Jan, Y.N., Jan, L.Y. & Berger, J.M. (2000) The polar T1 interface is linked to conformational changes that open the voltage-gated potassium channel. Cell 102, 657–670. Muto, A., Tashiro, S., Tsuchiya, H., Kume, A., Kanno, M., Ito, E., Yamamoto, M. & Igarashi, K. (2002) Activation of Maf/ AP-1 repressor Bach2 by oxidative stress promotes apoptosis and its interaction with promyelocytic leukemia nuclear bodies. J. Biol. Chem. 277, 20724–33. Nanao, M.H., Zhou, W., Pfaffinger, P.J. & Choe, S. (2003) Determining the basis of channel-tetramerization specificity by x-ray crystallography and a sequence-comparison algorithm: family Values (FamVal). Proc. Natl. Acad. Sci. USA 100, 8670–8675. Ochiai, K., Katoh, Y., Ikura, T., Hoshikawa, Y., Noda, T., Karasuyama, H., Tashiro, S., Muto, A. & Igarashi, K. (2006) Plasmacytic transcription factor Blimp-1 is repressed by Bach2 in B cells. J. Biol. Chem. 281, 38226–38234. Orlicky, S., Tang, X., Willems, A., Tyers, M. & Sicheri, F. (2003) Structural basis for phosphodependent substrate selection and orientation by the SCFCdc4 ubiquitin ligase. Cell 112, 142– 144. Otwinowski, Z. & Minor, W. (1997) Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326. Oyake, T., Itoh, K., Motohashi, H., Hayashi, N., Hoshino, H., Nishizawa, M., Yamamoto, M. & Igarashi, K. (1996) Bach proteins belong to a novel family of BTB-basic leucine zipper transcription factors that interact with MafK and regulate transcription through the NF-E2 site. Mol. Cell. Biol. 16, 6083– 6095. Perez-Torrado, R., Yamada, D. & Defossez, P.A. (2006) Born to bind: the BTB protein-protein interaction domain. BioEssays 28, 1194–1202. Pintard, L., Willis, J.H., Willems, A., Johnson, J.L., Srayko, M., Kurz, T., Glaser, S., Mains, P.E., Tyers, M., Bowerman, B. & Peter, M. (2003) The BTB protein MEL-26 is a substratespecific adaptor of the CUL-3 ubiquitin-ligase. Nature 425, 311–316.


Genes to Cells (2009) 14, 167–178

177

N Ito et al. Pintard, L., Willems, W.A. & Peter, M. (2004) Cullin-based ubiquitin ligases: Cul3-BTB complexes join the family. EMBO J. 23, 1681–1687. Schubot, F.D., Tropea, J.E. & Waugh, D.S. (2006) Structure of the POZ domain of human LRF, a master regulator of oncogenesis. Biochem. Biophys. Res. Commun. 351, 1–6. Stead, M.A., Trinh, C.H., Garnett, J.A., Carr, S.B., Baron, A.J., Edwards, T.A. & Wright, S.C. (2007) A b-sheet interaction interface directs the tetramerisation of the Miz-1 POZ domain. J. Mol. Biol. 373, 820–826. Stebbins, C.E., Kaelin, W.G.J. & Pavletich, N.P. (1999) Structure of the VHL-ElonginC-ElonginB complex: implications for VHL tumor suppressor function. Science 284, 455–461. Stogios, P.J., Downs, G.S., Jauhal, J.J.S., Nandra, S.K. & Prive, G.G. (2005) Sequence and structural analysis of BTB domain proteins. Genome Biol. 6, R82. Stogios, P.J., Chen, L. & Prive, G.G. (2007) Crystal structure of the BTB domain from the LRF/ZBTB7 transcriptional regulator. Protein Sci. 16, 336–342. Sun, J., Brand, M., Zenke, Y., Tashiro, S., Groudine, M. & Igarashi, K. (2004) Heme regulates the dynamic exchange of Bach1 and NF-E2-related factors in the Maf transcription factor network. Proc. Natl Acad. Sci. USA 101, 1461–1466. Tashiro, S., Muto, A., Tanimoto, K., Tsuchiya, H., Suzuki, H., Hoshino, H., Yoshida, M., Walter, J. & Igarashi, K. (2004) Repression of PML nuclear body-associated transcription by oxidative stress-activated Bach2. Mol. Cell. Biol. 24, 3473– 3484. Wilkins, A., Ping, Q. & Carpenter, C.L. (2004) RhoBTB2 is a

178

Genes to Cells (2009) 14, 167–178

substrate of the mammalian Cul3 ubiquitin ligase complex. Genes Dev. 18, 856–861. Willems, A.R., Schwab, M. & Tyers, M. (2004) A hitchhiker’s guide to the cullin ubiquitin ligases: SCF and its kin. Biochim. Biophys. Acta 1695, 133–170. Wu, G., Xu, G., Schulman, B.A., Jeffrey, P.D., Harper, J.W. & Pavletich, N.P. (2003) Structure of a b-TrCP1-Skp1-b-catenin complex: destruction motif binding and lysine specificity of the SCF(b-TrCP1) ubiquitin ligase. Mol. Cell 11, 1445–1456. Xu, L., Wei, Y., Reboul, J., Vaglio, P., Shin, T.H., Vidal, M., Elledge, S.J. & Harper, J.W. (2003) BTB proteins are substratespecific adaptors in an SCF-like modular ubiquitin ligase containing CUL-3. Nature 425, 316–321. Yoshida, C., Tokumasu, F., Hohmura, K.I., Bungert, J., Hayashi, N., Nagasawa, T., Engel, J.D., Yamamoto, M., Takeyasu, K. & Igarashi, K. (1999) Long range interaction of cis-DNA elements mediated by architectural transcription factor Bach1. Genes Cells 4, 643–655. Zheng, N., Schulman, B.A., Song, L., Miller, J.J., Jeffrey, P.D., Wang, P., Chu, C., Koepp, D.M., Elledge, S.J., Pagano, M., Conaway, R.C., Conaway, J.W., Harper, J.W. & Pavletich, N.P. (2002) Structure of the Cul1-Rbx1-Skp1-F boxSkp2 SCF ubiquitin ligase complex. Nature 416, 703–709. Ziegelbauer, J., Shan, B., Yager, D., Larabell, C., Hoffmann, B. & Tjian, R. (2001) Transcription factor MIZ-1 is regulated via microtubule association. Mol. Cell 8, 339–349. Received: 17 August 2008 Accepted: 4 November 2008


Crystal structure of the Bach1 BTB domain and its regulation of

Recommend Documents