Crystal structure of HutZ, a heme storage protein from Vibrio cholerae: A structural mismatch observed in the region of high sequence conservation
- First Online:
- Cite this article as:
- Liu, X., Gong, J., Wei, T. et al. BMC Struct Biol (2012) 12: 23. doi:10.1186/1472-6807-12-23
- 4.9k Downloads
HutZ is the sole heme storage protein identified in the pathogenic bacterium Vibrio cholerae and is required for optimal heme utilization. However, no heme oxygenase activity has been observed with this protein. Thus far, HutZ’s structure and heme-binding mechanism are unknown.
We report the first crystal structure of HutZ in a homodimer determined at 2.0 Å resolution. The HutZ structure adopted a typical split-barrel fold. Through a docking study and site-directed mutagenesis, a heme-binding model for the HutZ dimer is proposed. Very interestingly, structural superimposition of HutZ and its homologous protein HugZ, a heme oxygenase from Helicobacter pylori, exhibited a structural mismatch of one amino acid residue in β6 of HutZ, although residues involved in this region are highly conserved in both proteins. Derived homologous models of different single point variants with model evaluations suggested that Pro140 of HutZ, corresponding to Phe215 of HugZ, might have been the main contributor to the structural mismatch. This mismatch initiates more divergent structural characteristics towards their C-terminal regions, which are essential features for the heme-binding of HugZ as a heme oxygenase.
HutZ’s deficiency in heme oxygenase activity might derive from its residue shift relative to the heme oxygenase HugZ. This residue shift also emphasized a limitation of the traditional template selection criterion for homology modeling.
KeywordsHutZ Heme-binding Crystal structure Homology modeling
Iron is an essential element for the Gram-negative pathogenic bacterium Vibrio cholerae. It plays important roles in the microbe’s survival and its ability to cause the diarrheal disease cholera of V. cholerae. Nevertheless, the concentration of free iron is extremely low in the environment as well as in the human hosts. Under iron starvation conditions, V. cholerae has evolved several high-affinity iron uptake systems . Synthesis and secretion of the catechol-type siderophore vibriobactin is the main mechanism for obtaining iron . Siderophores, such as schizokinen , enterobactin [4, 5] and ferrichrome , produced by other microorganisms, can also be utilized by V. cholerae.
Heme, an excellent iron source in the environment and human hosts, can be used by V. cholerae in the free form or with heme-binding proteins [7, 8]. It is first transported into the cell with the assistance of the corresponding TonB-dependent outer membrane receptors and ATP-binding cassette transporter system proteins [3, 9], and then the iron released from heme by cytoplasmic heme oxygenase.
However, to date, no heme oxygenase has been reported in V. cholerae. BLAST searches against the NCBI database  have also returned no heme oxygenase homologues for V. cholerae. Therefore, the fate of heme after it enters V. cholerae’s cytoplasm remains mysterious, and little is known regarding which proteins contribute to heme utilization in the cytoplasm. In 2004, Wyckoff and coworkers identified in V. cholerae the heme-binding protein HutZ, a protein essential for the optimal utilization of heme as an iron source . However, no heme oxygenase activity has been observed for this protein, which indicates that HutZ serves only as a heme storage protein . NCBI BLAST searches have shown that HutZ belongs to the pyridoxine-5'-phosphate (PNP) oxidase-like superfamily and shares 35% sequence identity with the heme oxygenases HugZ from Helicobacter pylori[12, 13] and ChuZ from Campylobacter jejuni. Meanwhile, sequence comparisons have shown that HutZ shares low sequence identities (~10%) with other known representative heme oxygenases, such as HmuO from Corynebacterium diphtheria, HemO from Neisseria meningitidis, IsdG from Bacillus anthracis , and ChuS from Escherichia coli O157:H7 .
To gain further insight into the mechanism of the mysterious function of HutZ, this protein from V. cholerae (strain N16961) was overexpressed and its crystal structure determined at 2.0 Å resolution. The HutZ structure appeared to a typical split-barrel fold that is usually conserved in FMN-binding proteins. As heme did not cocrystallize with HutZ, molecular docking and site-directed mutation experiments were performed to investigate the interaction mechanism of HutZ and heme.
In the absence of experimentally solved protein structures, homology modeling is widely used to predict a structure based on one or more known structures of homologous proteins. Such a model is developed based on the general rule that similar protein sequences correspond to similar protein structures. Notably, in the comparison of the structures of HutZ and HugZ, a structural mismatch was identified in one amino acid residue after the corner of β6 of HutZ, where both protein sequences are identical except for two pairs of different amino acid residues. This observation suggested a potential hazard in the accuracy or application of the homology modeling method, in particular in the accuracy of the input sequence alignment. In the light of this, a series of homologous models were constructed to further verify the sequence-structure relationship between HutZ and HugZ and to explore functional implications from the HutZ structure. As ChuZ and HugZ share high sequence identity (53%) and structural similarity (root mean square deviation (RMSD) of 2.1 Å for the protein backbone), HugZ was employed as the representative heme oxygenase for homology modeling of different HutZ variants and for structural comparisons.
The overall structure of HutZ
Structural comparison between HutZ and HugZ
Structural modeling of HutZ and its variants
Anfinsen’s dogma  states that a protein’s native structure is determined by its amino acid sequence and is a stable and kinetically accessible minimum of free energy. Accordingly, the local structural divergence between HutZ and HugZ should result from certain amino acid diversities. Therefore five homologous models for HutZ and its variants were constructed, and their rationalities were assessed to determine their essential residues. The compatibility of every amino acid with its conformation was assessed by ProQres, but the focus was on the β6 of HutZ and the corresponding regions in other models, where the structural mismatch occured.
Next, the goal was to identify which residue was the main contributor to HutZ’s structural shift. From the beginning of the corner region (Pro140), where the structural mismatch begins, to Tyr153 of HutZ, there were a total of four different residue pairs between HutZ and HugZ (Figure 4A). These residues were very likely responsible for the structural shift and, thus, four single point variants (P140F, E141K, Q142E and L144R) for HutZ were designed and four models (M2-M5) generated for them, respectively (Figure 6). The mutated residues in HutZ were simply replaced with the corresponding residues of HugZ, and, as with M1, all variant models used the HugZ structure as a template, such that all of them were aligned with HugZ in one-to-one correspondence both on the sequence and structural levels. Structural evaluations of these models suggested three things (Figure 6). First, the two crystal structures were most favored by their sequences, as they generally received higher scores than the homologous models. Second, the score curves of the HugZ crystal structure and all homologous models exhibited similar shapes, which were obviously different from the HutZ crystal structure, because all models were constructed using the HugZ crystal structure as a template. Last, among all homologous models, M2 was evaluated as the best, while M1, M3, M4 and M5 were close to one another, suggesting that the sequence of the variant P140F better fitted the HugZ structure than those of other variants. In other words, although the residues Pro140, Glu141, Gln142 and Leu144 may together have contributed to the formation of HutZ’s structural shift, Pro140 played a more important role than other residues. Moreover, Pro140 was located at the very beginning of the corner region, where the structural shift also began, underlining the role of Pro140 in the local folding of HutZ β6.
HutZ, a unique heme storage protein identified in V. cholerae, is necessary for optimal heme utilization [11, 20], but no heme oxygenase activity has been detected for HutZ. The crystal structure of the heme oxygenase HugZ shows that there are two symmetric active sites located at the HugZ dimmer interface, formed by the C-terminal region (β8-β11 and the C-terminal loop) from one monomer and α7 from the other . The C-terminal loop functions as a flexible portion of the active site, which is supposed to keep the substrate heme molecule in proper conformation for the heme oxygenase activity as well as to close off the active pocket . In the HutZ structure, the homologous heme-binding pocket involved β5 and β6, which corresponded with β8 and β9 in HugZ. The structure from β7 to the C-terminal loop, which corresponded to β10, β11 and the C-terminal loop in HugZ, was truncated in the present structure. Thus, it was unclear whether the missing portion participated in the heme-binding pocket or it resulted in the lack of HutZ enzymic activity. His170 from the C-terminal of HutZ, corresponds to the fully conserved His245 in HugZ that is responsible for the heme iron atom coordination . Notably here, HutZ’s His170 did not contribute to iron coordination and enzymatic activity in HutZ as, when mutant H170A was reconstituted with heme, the Soret peak did not change in comparison with the native HutZ-heme complex (data not shown). This study of the structural mismatch between HutZ and HugZ provided an important clue for resolving this issue. Because the corner of β6 of HutZ includes one more amino acid residue than that of HugZ, HutZ’s β6-2 (five residues) ends one amino acid earlier than HugZ’s β9-2 (six residues). It is known that the torsion angle of N-Cα-C-N in the backbone of a β-strand is about 120°. Therefore, the main chains of the end residues of the two β-strands (β6-2 and β9-2) point in different directions to form a 60° angle (Figure 4D). This direction deviation led to an absolute mismatch between the structures of HutZ (Phe149 and Gly150) and HugZ (Gly223 and Phe224) \(Figure 4C), and might have led to even larger structural differences in their C-terminal regions (missing in the present structure). In this regard, it was speculated that the C-terminal loop, which is essential for HugZ enzymatic activity, might have turned away from the heme-binding pocket of HutZ and yielded the pocket more exposed than that of HugZ. This might then have resulted in low heme-binding affinity and a deficiency in HutZ enzymatic activity. In addition, the possibility was excluded here that truncating the protein resulted in a structural shift of β6-2 in HutZ for three reasons. First, HutZ was truncated here because of the twinning and fragile crystals of the full length protein. If the complete HutZ structure matched HugZ at β6-2 and the C-terminal loop, good full length crystals would have been obtained, as have been attained by Hu and coworkers . Second, an extensive hydrogen bond network connected two parallel β-stands and maintained their stable conformations. Truncating the residues that follow β6-2 might not have provided enough energy to destroy the normal hydrogen bonding and rearrange them. And third, the hypothetical HutZ structure (M1) that matches HugZ at β6-2 evaluated even more poorly than the present truncated HutZ structure.
Protein expression, purification and site-directed mutagenesis
The hutZ gene was amplified from genomic DNA of V. cholerae and subcloned into the pET21b expression vector (Novagen, EMD Biosciences, Inc., Darmstadt, DE) between the NdeI and XhoI restriction cut sites. The N-terminal and C-terminal regions (residues 1–12 and 151–176, respectively) were removed during cloning, resulting in a construct of 13-150-His tagged fusion protein for further expression.
BL21 (DE3) cells containing plasmids for the recombinant HutZ protein were grown in L Broth media supplemented with 100 μg/mL ampicillin. Once the culture attained an OD600 of 1.0, the incubation temperature was decreased to 15°C, and protein expression induced by adding isopropyl β-D-1-thiogalactopyranoside to a final concentration of 0.10 mM. After 8 h of expression, the cells were harvested by centrifugation and the pellet resuspended in lysis buffer (15 mM Tris–HCl, pH 8.0, 100 mM NaCl, and 1 mM phenyl methane sulfonyl fluoride), and lysed by sonication. Cell debris was then removed by centrifugation at 28500 × g for 45 min. The resulting soluble fractions containing recombinant protein HutZ was loaded onto a Ni-chelating Sepharose affinity column (GE Healthcare, Buckinghamshire, UK) equilibrated with lysis buffer. The affinity column was washed extensively with lysis buffer and all proteins eluted with elution buffer (20 mM Tris–HCl, pH 8.0, 50 mM NaCl and 200 mM imidazole). The elution protein was further purified using an ion-exchange column (Source 15Q, GE Healthcare), conditioned with equilibration buffer (25 mM Tris–HCl, pH 8.0, 3 mM DTT), and eluted using a linear 150 mL gradient of 0–0.5 M NaCl. Finally, HutZ protein was purified using size exclusion chromatography (Superdex-200, GE Healthcare). Fractions were pooled according to protein purity monitored by SDS-PAGE, and the final protein concentration at 12 mg/mL. All mutant proteins were purified using the same procedure.
Crystallization and data collection
HutZ was crystallized using the sitting drop vapor diffusion method at 20°C by mixing equal volumes of protein with reservoir solution containing 1.6 M Na/K phosphate and 0.1 M HEPES at pH 7.5. Because the crystals of full length HutZ were twinning and fragile and could not be used to obtain high resolution diffraction data, the protein was truncated to amino acid residues 13–150 to grow high quality crystals, which appeared about seven to ten d and reached full size in two wk. Diffraction data were collected at the Shanghai Sychrotron Radiation Facility beamline BL17U1. To prevent radiation damage, crystals were transferred to a cryoprotectant buffer containing 15% glycerol (v/v) plus reservoir buffer and then flash-cooled using a nitrogen stream, with the temperature around the crystals maintained at 100 K throughout data collection. Data sets were processed using the HKL2000 software suite . The crystals belonged to the space group P41 with four macromolecules in an asymmetric unit, with unit cell dimensions of a = b = 80.1 Å and c = 125.8 Å.
Statistics of crystallographic analysis
unit cell (Å)
Bond lengths (Å)
Bond angles (°)
Ramachandran plot (%)d
Most favored (%)
Additionally allowed (%)
Generously allowed (%)
Docking and homology modeling
AutoDock 4.2  was used to perform flexible docking in the HutZ-heme complex. The heme was restricted within a grid box (50 × 50 × 30 points in dimension and 0.375 Å spacing) that enveloped the HutZ binding cleft, which was determined through similarities with the known crystal structure of the HugZ-heme complex. Docking searches were executed using the Lamarckian genetic algorithm and a maximum number of 25,000,000 energy evaluations. After docking searches were finished, the first-ranked model, based on binding energy, was selected as the final result from 50 candidate solutions.
A series of homologous models of HutZ and its variants were generated using the MODELLER 9.9 program package  with HugZ (PDB code: 3GAS) as a template. The input files for each target sequence were a pairwise sequence alignment (template and target) and the coordinate file of the template. The number of output models for each target was set to five, and other MODELLER options at default. The qualities of all resulting models were evaluated with ProQres , a neural-network based local protein model quality predictor, which analyzes the compatibility of every amino acid residue with its conformation in a three-dimensional model.
We determined the crystal structure of HutZ from V. cholerae at 2.0 Å resolution and compared the structure with that of its homologous protein HugZ. The structural mismatch between HutZ and HugZ presented a rare case in which the structural alignment was not in accordance with the sequence alignment for two highly similar proteins (structure RMSD = 1.645 Ã and sequence identity = 35%). This observation suggested a potential hazard in the assumed accuracy of template selection of the traditional homology modeling method. If a homologous model of HutZ is constructed using the default sequence alignment with HugZ and the HugZ structure as a template, the resulting model will be inaccurate.
The atomic coordinates and structure factors have been deposited in the Protein Data Bank (http://www.rcsb.org/pdb/) with the accession code 3TGV.
The genomic DNA of Vibrio cholerae is a gift from Prof. Bonnie Bassler. The authors thank the staff at beamline BL17u1 at the Shanghai Synchrotron Radiation facility for support with data collection. This work was supported by State Key Laboratory of Microbial Technology (SDU), the Grant of Hi-Tech Research and Development Program of China (2006AA02A324) and Promotive Research Fund for Excellent Young and Middle-aged Scientists of Shandong Province (BS2012SW006).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.