|Home | About | Journals | Submit | Contact Us | Français|
We have recombinantly expressed and purified the ligand binding domains (LBDs) of four insect nuclear receptors of the E75 family. The Drosophila melanogaster and Bombyx mori nuclear receptors were purified as ferric hemoproteins with Soret maxima at 424 nm, whereas their ferrous form had a Soret maximum at 425 nm that responds to ·NO and CO binding. In contrast, the purified LBD of Oncopeltus fasciatus displayed a Soret maximum at 415 nm for the ferric protein that shifted to 425 nm in its ferrous state. Binding of ·NO to the heme moiety of D. melanogaster and B. mori E75 LBD resulted in the appearance of a peak at 385 nm, whereas this peak appeared at 416 nm in the case of the O. fasciatus hemoprotein, resembling the behaviour displayed by its human homolog Rev-erbβ. HPLC analysis revealed that, unlike the D. melanogaster and B. mori counterparts, the heme group of O. fasciatus is covalently attached to the protein through the side-chains of two amino acids. The large sequence homology with O. fasciatus E75 led us to clone and express the LBD of Blattella germanica, which established that its spectral properties closely resemble those of O. fasciatus and that it also has the heme group covalently bound to the protein. Hence, ·NO/CO regulation of the transcriptional activity of these nuclear receptors might be differently controlled among various insect species. In addition, covalent heme binding provides strong evidence that at least some of these nuclear receptors function as diatomic gas sensors rather than heme sensors. Finally, our findings expand the classes of hemoproteins in which the heme group is normally covalently attached to the polypeptide chain.
Nuclear receptors, the largest superfamily of transcription factors, are ligand-regulated polypeptides that share common domain architecture. Binding of small molecules to the ligand binding domain (LBD) of nuclear receptors modulates their association to specific DNA motifs. In insects, the early-induced gene E75 has been implicated genetically in repression of several genes in the ecdysone-triggered cascade (1) and it is well-established that E75 also acts as a repressor in transient transfection assays (2–4). Heterodimerization of E75 with HR3 or SMRTER is known to block their ability to activate transcription (3, 5, 6). Interestingly, the Drosophila melanogaster E75 nuclear receptor, and more specifically its LBD, is a protein module that binds heme and diatomic gases such as ·NO and CO (3). Full-length D. melanogaster E75 is isolated as a heme-bound ferric hemoprotein when obtained from the insect pupae and its isolated LBD is also a hemoprotein when recombinantly expressed and purified from both bacteria and baculovirus-infected cells (3, 7, 8).
From a functional point of view, coordination of diatomic gases to the heme moiety of D. melanogaster E75 induces a conformational change that interferes with its interaction with HR3 and SMRTER nuclear receptors, discontinuing repression of their transcriptional activity (3, 6). Binding of ·NO to D. melanogaster E75 displaces the Cys and His axial ligands rendering a pentacoordinate Fe(II)NO hemoprotein and justifying the observation that ·NO functions as an antagonist of E75 repressor activity (3, 5, 6, 8). In flies, nitric oxide synthase-derived ·NO cancels the repression exerted by E75 on HR3 in the prothoracic gland (5). Furthermore, during the transition of larvae to prepupae, ·NO can almost completely inhibit the repressor activity of E75 by preventing its recruitment by SMRTER (6). Finally, CO binding to purified D. melanogaster E75 LBD produces a hexacoordinate Fe(II)CO complex with a sixth neutral ligand (7, 8), stabilizes the protein, and abrogates its interaction with an HR3 peptide (3) although it is not known if, in vivo, ferrous CO-bound D. melanogaster E75 LBD fails to bind to other nuclear receptors such as HR3 or SMRTER.
Within mammalian LBDs, the closest homologs of insect E75 are nuclear receptors Rev-erbα (NR1D1) and Rev-erbβ (NR1D2), two related proteins that generally act as transcriptional repressors, either on their own or recruiting co-repressor proteins (9). In most cases, Rev-erbs are present in the nucleus and bind as monomers to the Rev responsive element or as dimers to a Rev-RE direct repeat, RevDR-2 (10). As in the case of insect E75 LBDs, both Rev-erbα and Rev-erbβ bind heme and the reversible binding of heme seems to regulate Rev-erbβ transcription activity (11–13). Whereas wild-type Rev-erbβ can repress transcription of target genes, its His602Phe mutant, which is unable to bind heme, does not display this transcriptional repressor activity (12). Since heme cellular levels vary during the circadian cycle, Rev-erbα and Rev-erbβ are considered to link the circadian cycle and metabolism (14). Initially, it was proposed that the transcriptional activity of Rev-erbα or Rev-erbβ was not affected by reduction of the heme iron with sodium dithionite or by the addition of ·NO donors (12), although recent data suggest that both recombinant Rev-erbα and Rev-erbβ can indeed bind ·NO and CO (8, 13, 15). Moreover, heme-saturated Rev-erbβ can also bind ·NO, hence regulating the binding to co-repressors (13).
In this manuscript we have characterized the properties of four insect E75 LBDs. Purified recombinant E75 LBD from the silkworm Bombyx mori significantly resembles that of Drosophila melanogaster. Unexpectedly, we have observed that the heme moiety becomes covalently attached to large milkweed bug (Oncopeltus fasciatus) and German cockroach (Blattella germanica) E75 LBDs yielding proteins with spectral and functional properties that resemble those of Rev-erbα and Rev-erbβ rather than those of their D. melanogaster and B. mori counterparts. In hemoproteins, covalent attachment of the heme moiety to the protein is not without precedent. In cytochrome c, the heme moiety is covalently attached to the protein through thioether bonds between its two vinyl groups and the side-chains of two Cys residues (16). In the CYP4 family of cytochrome P450 enzymes, covalent attachment of the heme 5-methyl group to the side-chain of a Glu residue in helix I via an ester link has been observed in CYP4A, CYP4B and CYP4F proteins (17–21). In the case of lactoperoxidase and eosinophil peroxidase, the side-chains of a Glu and an Asp form ester links with the heme 1- and 5- methyl groups (22), respectively, whereas in myeloperoxidase, two types of linkages are found, one involving an ester link between the side-chains of Glu and Asp residues and the 1- and 5-methyl groups of heme, respectively, and the other a sulfonium link between the sulphur of a Met residue and the β-carbon of the heme 2-vinyl group (23). In cytochrome c peroxidase and ascorbate peroxidase, covalent linkages between the side-chain of a Trp residue and one of the vinyl groups of the heme moiety have been observed (24, 25).
Both the covalent attachment of heme and the various results reported herein have been interpreted in the light of the heme, CO, ·NO and redox sensing properties of this family of nuclear receptors.
Buffers, chemicals, oligonucleotides, hemin and common laboratory reagents were obtained from Sigma-Aldrich if not otherwise indicated. Pfu polymerase, T4 DNA ligase, restriction endonucleases and Molecular Mass markers were obtained from Fermentas. Ni-NTA resin was from Qiagen. CO and ·NO gases were from Airgas. The RP523 E. coli strain deficient in heme biosynthesis was obtained from the Yale University genetic stock center. Fe(III) mesoporphyrin IX was purchased from Livchem (Mannheim, Germany).
Drosophila melanogaster E75 LBD was retrieved by PCR from a cDNA preparation of fly embryos using oligonucleotides that matched the published sequence. The E75B LBD from Oncopeltus fasciatus was amplified by PCR using the complete cDNA extracted from embryos (18, 22, 26, 30 and 34 h) and was a gift from Dr. Deniz F. Erezyilmaz (Princeton University, USA) (26). The Bombyx mori E75A cDNA was a generous gift from Dr. Kostas Iatrou (Athens, Greece) (4). The cDNA of Blattella germanica E75A was provided by Dr. David Martin Casacuberta (Barcelona, Spain) (27).
In all cases we used the NdeI (5´ end) and XbaI (3´ end) of the bacterial expression vector pCWori in which a hexa-His tag has been previously introduced in frame at the N-terminus end of the recombinant protein as previously described (28, 29). The DNA encoding D. melanogaster E75 LBD was amplified by PCR. A forward oligonucleotide (5´-ACCCAGAATCGCGGCCAGCAGCGAGCC-3´) and a reverse oligonucleotide (5´-GGGCGACTTGTTCTGCTGGCCATCGCTGTTG-3´) were first used to amplify the LBD (residues 341–604). A second PCR with Pfu polymerase using this template was performed using the oligonucleotides (5´-ACCCAGCATATGGGCCAGCAGCGAGCC-3´, NdeI site underlined) and (5´-, GGGCGACTTTCTAGACTAGCCATCGCTGTTG-3´, XbaI site underlined). Likewise, the DNA encoding the E75B LBD from O. fasciatus was amplified by PCR using the complete cDNA extracted from O. fasciatus embryos. A forward oligonucleotide (5´-GCAGAGCACCAACTCCAAGTGCCAGGAG-3´) and a reverse oligonucleotide (5´-CATTGAACCCCACATTTCGTGTTGCTG-3´) were first used to amplify the LBD (residues 88–348). A second PCR with Pfu polymerase was performed using the oligonucleotides (5´-GCAGAGCACCAACCATATGTGCCAGGAG-3´, NdeI site underlined) and (5´-CATTGAACTCTAGATTTCGTGTTGCTG-3´, XbaI site underlined). Site-directed mutagenesis was performed following the QuikChange mutagenesis protocol. The O. fasciatus Glu158 codon was changed for Lys and the Met245 codon for Thr.
After a Clustal comparison between D. melanogaster and O. fasciatus amino acid sequences (Fig. S2) we selected a stretch -TLLKAG- approximately in the middle of helix 5 in the LBD that was identical in both proteins. A silent AflII was introduced at this position by site-directed mutagenesis in both the D. melanogaster and O. fasciatus cDNA. Chimera Dros/Onc consisted in the combined N-terminus of D. melanogaster E75 LBD and the C-terminus of O. fasciatus E75. Chimera Onc/Dros consisted in the N-terminus of O. fasciatus E75 LBD and the C-terminus of D. melanogaster E75. The two chimeras were referred to as chimera Dros/Onc (residues 341–446 of the D. melanogaster LBD followed by residues 194–348 of its O. fasciatus counterpart) and chimera Onc/Dros (residues 88–199 of the O. fasciatus LBD followed by residues 440–604 D. melanogaster counterpart). Both chimeric proteins were cloned and expressed in the pCWori vector.
Recombinant protein expression was performed in the E. coli protease-deficient strain BL21(DE3). Cells were grown in 2XYT medium in the absence of hemin, to an OD600 of ~1 prior to induction for 18 h at 25°C by the addition of 1 mM IPTG. After centrifugation, the cell pellet was resuspended in 100 mM Tris-HCl (pH 7.0) plus lysozyme and protease inhibitors and lysed by sonication. All the E75 proteins expressed in E. coli were purified by Ni-NTA affinity column chromatography (Qiagen). The column was extensively washed with 200 ml of 100 mM Tris-HCl, pH 7.0 plus 500 mM NaCl, followed by a 100 ml wash with 100 mM Tris-HCl, pH 7.0, 250 mM NaCl, 30 mM imidazole followed by a final 100 ml wash with 100 mM Tris-HCl, pH 7.0, 250 mM NaCl, 45 mM imidazole. The bound E75 proteins were eluted with 25 mM Tris-HCl, pH 7.0 plus 200 mM imidazole. Pooled fractions were dialyzed in 25 mM Tris-HCl, pH 7.0, 100 mM NaCl buffer. Next, we loaded the eluted protein in a gel filtration Superdex 75 preparative grade column (GE Healthcare) equilibrated and eluted with 50 mM Tris-HCl, pH 7.0, 100 mM NaCl, in a FPLC System (Pharmacia Biotech, Uppsala, Sweden). Coloured fractions containing pure E75 were identified by SDS-PAGE. The proteins were flash frozen in liquid nitrogen.
In order to obtain the ferric, ferrous, ferrous-CO and ferrous-NO absorbance spectra for various insect E75 LBDs, samples (3–5 µM) were prepared under anaerobic conditions inside a glove box. Spectra were collected with protein samples (700 µl) prepared in 25 mM Tris, pH 7.0, 100 mM NaCl degassed buffer were purged of oxygen by flowing argon through the headspace of a septum-sealed vial for ~5 min. Reduction of Fe(III) protein samples using sodium dithionite was performed under an atmosphere of nitrogen using the addition of an anaerobically prepared stock solution of sodium dithionite to a final concentration of 1 mM. CO or ·NO adducts were prepared by the injection of CO(g) or ·NO(g) into the headspace of the septum-sealed cuvette via a gastight syringe followed by gentle agitation of the sample. In general, 100–200 µl of CO or ·NO gas were injected to render a complete conversion to the respective adduct. The absorption spectra at room temperature were collected on a Cary 50 Bio UV-visible spectrophotometer (Varian) until no further change was observed.
This assay allows the measurement of heme concentration, but also allows to determine if the prosthetic group of the protein is heme a, heme b or heme c. A stock pyridine solution containing 400 mM NaOH and 34% (v) pyridine was prepared in water. Then, 10 to 50 µl of concentrated protein were mixed with pyridine solution to a final volume of 500 µl in a 0.5-ml cuvette, and the oxidized spectrum was recorded between 500 and 600 nm (baseline). A few grains of solid sodium dithionite (2–5 mg) were then added and the sample was mixed by flipping the cuvette sealed with a parafilm; several successive spectra of the reduced pyridine hemochromes were recorded on a Beckman DU 640 spectrophotometer (Fullerton, CA, USA).
CD spectra were recorded on a Jasco J-715 spectropolarimeter (Jasco Inc., Easton, MD, USA) using a cuvette path length of 0.1 cm and spectral collection in the range of 250–700 nm at 25 °C. The buffer used was 20 mM Tris, pH 8.5. A minimum of five spectra were accumulated for each sample and the contribution of the buffer was always subtracted. Raw ellipticity data were converted to mean residue ellipticity before plotting.
Samples were analyzed by direct injection of 100 µg of protein onto a 250 × 4.6-mm Beckman Coulter Ultrasphere C18 reversed phase column on a Beckman Coulter HPLC instrument. The protein was eluted with a linear gradient of 25–80% acetonitrile in water (0.1% trifluoroacetic acid) over 60 min (1 ml/min) with detection at 214 and 400 nm. As a control, a stock solution of hemin was prepared by dissolving 3 mg in 100 µl of 0.1 M NaOH, followed by the addition of 900 µl of water and filtrated with a filter of 0.22 µm. A 10 µl stock aliquot was diluted into 240 µl of water (0.1% trifluoroacetic acid) for analysis.
For the analysis of heme-containing peptides, E75B LBD from O. fasciatus (30 µl of a 4.5 mg/ml solution in 20 mM Tris pH 8.5) was digested with trypsin (20 µg) in 220 µl of 0.09 mM Tris pH 7.0, for 3 h at 37 °C. After this partial digestion, 50 µg of digested protein were chromatographed by direct injection onto a 250 × 4.6-mm Beckman Coulter Ultrasphere C18 reversed phase column on a Beckman Coulter HPLC instrument, with a linear gradient of 30–60% acetonitrile in water (0.1% trifluoroacetic acid) over 20 min (1 ml/min) with detection at 214 and 400 nm. A mixture of heme-containing peptides eluting at different times were collected, vacuum concentrated and analyzed by mass spectrometry. The E75B LBD from O. fasciatus (37.5 µl of a 4 mg/ml solution in 50 mM Tris pH 7, 100 mM NaCl) was also digested with proteinase K (18.75 µg) for 20 min at 37 °C. After this partial digestion, 50 µg of digested protein were chromatographed and analyzed as previously described.
Continuous-wave (CW) EPR spectra were recorded at a low temperature (10°K) with an X-band EMX Bruker spectrometer equipped with an Oxford Instruments ESR 900 helium flow cryostat. Before being frozen, 20% (v/v) glycerol as a cryoprotectant was added to all the samples. Samples were introduced in a synthetic Suprasil quartz EPR tube and vacuum sealed before analysis.
1 µl sample was spotted onto a MALDI target plate and allowed to air-dry at room temperature. Then, 0.5 µl of a 3 mg/ml of α-cyano-4-hydroxycinnamic acid matrix (Sigma-Aldrich) in 50% acetonitrile were added to the dried spots and allowed again to air-dry at room temperature. MALDI-TOF MS analyses were performed in a 4800 Proteomics Analyzer MALDI-TOF/TOF mass spectrometer (Applied Biosystems, MDS Sciex, Toronto, Canada) at the Genomics and Proteomics Center, Complutense University of Madrid. The MALDI-TOF/TOF operated in positive reflector mode with an accelerating voltage of 20000 V. Selected peptides, were subjected to MS/MS sequencing analyses using the 4800 Proteomics Analyzer (Applied Biosystems, Framingham, MA). Suitable precursors from the MS spectra were selected for MS/MS analyses with CID on (atmospheric gas was used) 1 Kv ion reflector mode and precursor mass Windows +/− 4 Da. The plate model & default calibration were optimized for the MS/MS spectra processing. De novo sequencing from fragmentation spectra of peptides was performed using DeNovo tool software (Applied Biosystems), tentative sequences were manually checked and validated.
The pCWori plasmid was used to transform the RP523 E. coli strain and expression was performed aerobically using 250 ml 2XYT cultures. When protein expression was induced by 1 mM IPTG addition, 1.25 mL of a 6 mg/ml solution of either Fe(III) protoporphyrin IX (hemin) or Fe(III) mesoporphyrin IX (mesoheme) was added to the cultures as previously described (30). Expression was performed at 25 °C and purification of the recombinant hemoproteins was performed as described above.
All three insect E75 LBDs were cloned in the bacterial expression system pCWori, which is suitable for recombinant heme protein expression (28). Purification of the hexa-His tagged LBDs was performed by Ni-NTA chromatography followed by gel filtration in a Superdex H75 column. The LBDs of all three E75 nuclear receptors were soluble and purified to homogeneity as hemoproteins in the absence of hemin supplementation of the cultures (Fig. S1). Whereas both the D. melanogaster and B. mori LBDs were bright red and a Soret maximum centered at 424 nm with α/β bands at 574/543 nm, the O. fasciatus LBD was red-brownish with a Soret displaced to 415 nm, an α band almost non-existent and a β band centered at 532 nm (Fig. 1A). Electronic absorption spectra of the D. melanogaster and B. mori insect nuclear receptor LBD constructs suggest strongly that they contain a low-spin hexacoordinate Fe(III) heme in which a histidine and a cysteine (thiolate) are present as axial ligands, in agreement with previous studies using the D. melanogaster isoform (3, 7, 8). In contrast, the O. fasciatus LBD spectrum is consistent with a myoglobin-like high-spin hexacoordinate Fe(III) heme (31) in which a histidine and (very likely) a water molecule are present as axial ligands (Fig. 1A). In addition, the O. fasciatus protein displayed a clear CT1 band at 655 nm which is indicative of the presence of a high-spin Fe(III) (7, 32).
Next, we compared the absorbance spectra of the three hemoproteins with their circular dichroism spectra at identical wavelengths. The δ band (porphyrin π-π* transitions) observed at 360 nm in the absorbance spectra of D. melanogaster and B. mori LBDs resulted in a positive CD band as well whereas this band was barely distinguishable from the O. fasciatus spectra. The γ band (Soret) observed at 424 nm in the absorbance spectra of D. melanogaster and B. mori LBDs and 415 nm in the spectrum of their O. fasciatus counterpart resulted in a positive CD band in all cases, although the former two proteins also display a clear negative CD band at 440 nm that is absent from the O. fasciatus CD spectrum (Fig. 1B). The α/β bands of the D. melanogaster and B. mori LBDs also resulted in a positive signal in their CD spectra although the O. fasciatus CD spectrum lacks a clear signal coming from its β band. Interestingly, the appearance of the aforementioned CT1 band at 655 nm in the O. fasciatus absorbance spectrum is reflected as well by a negative CD band.
The electronic absorption spectrum obtained after reduction of the heme iron of D. melanogaster E75 LBD with sodium dithionite exhibited a Soret maximum at 425 nm and α/β bands at 559 and 530 nm. This slight red shift upon iron reduction is also accompanied by a sharpening of the α/β bands (Fig. 2A). Binding of CO to reduced D. melanogaster E75 LBD results in a Soret maximum at 420 nm and α/β bands at 569 and 539 nm, consistent with a hexacoordinate heme with a His side-chain (or other neutral donor) as axial ligand (Fig. 2A). Formation of a ·NO complex (dashed line) in D. melanogaster E75 LBD results in the appearance of a broad peak at 385 nm, consistent with a pentacoordinate NO-heme. This blue-displacement of the Soret with concomitant broadening of the α/β bands is similar to the pentacoordinate Fe(II)NO complex observed in RrCooA or soluble guanylate cyclase (33, 34). Significantly, reduction of the heme iron in O. fasciatus E75 LBD with sodium dithionite results in a blue shift of the Soret to 425 nm accompanied by a clear sharpening of the peak as well as a better defined α/β bands at 557 and 530 nm (Fig. 2B). In analogy with its D. melanogaster E75 homolog, the O. fasciatus CO complex (dotted line) shows a sharp Soret at 420 nm, albeit the α/β bands are of a similar intensity and marginally red-shifted to 566 and 533 nm. In contrast, the ·NO adduct of ferrous O. fasciatus E75 LBD (dashed line) is significantly blue-shifted when compared with the D. melanogaster counterpart, with a Soret maximum at 416 nm and the near disappearance of the α/β bands. This spectrum is clearly reminiscent of that of the ·NO adducts observed in the case of myoglobin (35), neuroglobin (36) or ChCooA (33), in which the heme is hexacoordinated with ·NO and a His residue side-chain as the iron ligands. Unexpectedly, the spectrum of the ·NO adduct of ferrous O. fasciatus E75 LBD, although strikingly different from the one observed for its D. melanogaster counterpart, is analogous to that found in its mammalian homolog Rev-erbβ (8, 13). Finally, the ferrous form of the B. mori E75 LBD displays a Soret maximum at 425 nm (somewhat obscured by sodium dithionite absorption) and sharp α/β bands at 559 and 530 nm (Fig. 2C). The spectra of the CO and ·NO Fe(II) complexes of the B. mori E75 LBD are almost indistinguishable from those of D. melanogaster E75 LBD, with Soret maxima at 421 and 385 nm respectively, albeit significantly different from those O. fasciatus E75 LBD.
We have also analyzed the three insect E75 LBDs in their ferric state by EPR in order to characterize and assign the heme iron axial ligands (Fig. 3). The EPR spectrum of D. melanogaster E75 LBD reflects heterogeneity in iron coordination, in agreement with previous data (7, 8), and contains at least two sets of rhombic signals associated with g values corresponding to low-spin ferric iron. No high-spin EPR signals were detected. All g values are within the range found for thiolate hemoproteins in which the cysteine ligand determines the EPR properties of the heme iron. A comparison of the g values with published data indicates that one component (g values: 2.54, 2.26, 1.87) corresponds to an N-donor iron ligand such as histidine and the other component (2.33, 2.26, 2.04), is compatible with O- or S-donors. The EPR spectra of B. mori and D. melanogaster E75 LBDs are very similar, although the former contains a third component with the following g values: E1 (2.54, 2.26, 1.87) and E2 (2.44, 2.26, 1.91), compatible with N- or O-donors, together with an E3 component (2.33, 2.26, 2.04). In contrast, the EPR spectrum of the O. fasciatus E75 LBD displays a remarkable high-spin EPR signal with a g value of 4.32 and a small high-spin EPR signal with a g value 6.02. Although the different rhombic signals are not properly defined, it can be inferred that g values are similar to those detected for the ferric form of Rev-erbβ LBD (8), the mammalian homologue of insect E75. The estimated contribution of each component to each spectrum is shown in Table 1. In addition, the tetragonal field (Δ/λ) and rhombicity (V/Δ) parameters associated with E1–E3 are also shown.
Next we analyzed by HPLC the elution profiles of D. melanogaster, O. fasciatus and B. mori E75 LBDs from a reversed phase C18 column. Elution using an acetonitrile gradient revealed that the free heme peak could be detected at ~42 min in the 400 nm channel (Fig. 4A). Whereas almost all of the D. melanogaster and B. mori heme moiety eluted as free heme with the polypeptide chain eluting at higher acetonitrile concentrations, the greater part of the O. fasciatus heme moiety appeared to be associated covalently with the protein (Fig. 4B, C and D). Thus, the heme moiety of O. fasciatus E75 LBD not only displayed distinct spectroscopic properties and ·NO binding, but is also attached covalently to the protein.
Although the formation of a pyridine hemochrome complex is a widely used method for the determination of holoprotein concentration in hemoproteins, it is also a very useful method for determining if the protein binds heme b. As expected, the pyridine hemochrome spectra of the D. melanogaster and B. mori E75 LBDs display maxima at 556 nm, characteristic of heme b (37) whereas formation of a pyridine hemochrome complex with the O. fasciatus E75 LBD gives rise to a maximum at 551 nm (Fig. 5). This complex is reminiscent of the pyridine hemochrome complex observed in the case of cytochrome c, a hemoprotein with heme c covalently bound (37).
Mass spectrometric analysis of proteins with heme covalently attached can sometimes reveal the amino acid side-chain involved in the linkage, at least in cases in which the proteolytic digestion does not alter the integrity of the hemopeptides. When we digested purified Oncopeltus fasciatus E75 LBD with trypsin and analysed the digestion products by MALDI-TOF a high coverage could be attained (Fig. S5). Unfortunately, we were unable to identify masses that could be assigned to hemopeptides with mass gains of 616 or 633 Da over the expected proteolytic trypsin products. Intriguingly, we observed peptides of 1474.88 Da and 1697.91 Da that, after fragmentation, rendered MS/MS spectra consistent with a 240FLMDSMFDFAER251 sequence in which Met245 had lost 34 Da or gained 189 Da, respectively (Fig. S6). These results led to the conclusion that the heme moiety might be attached to the polypeptide chain of the O. fasciatus E75 LBD through a methionine vinyl-sulfonium bond in analogy with that observed in the case of myeloperoxidase (23, 38) or perhaps through a methionine ethyl-sulfonium as in the Ser160Met mutant of ascorbate peroxidase (39). Hence, fragmentation of the Met245 side-chain after trypsin hydrolysis rendered a peptide 240FLMDSMFDFAER251 with a 34 Da mass loss. Likewise peptide 240FLMDSMFDFAER251 with a mass gain of 189 Da in Met245 very likely arises from a fragmented heme adduct that remains attached to the amino acid side-chain. Since no atomic information is available in the case of insect E75 LBDs, we inspected the position of Met486 in the recently obtained crystal structure of its mammalian homologue Rev-erbβ (13). As shown in Fig. S7, Met486 in Rev-erbβ would be the equivalent to O. fasciatus Met245 (Fig. S2) and both would be positioned in helix 7 with the sulphur atom in close proximity to one of the heme vinyl groups, hence reinforcing the mass spectrometric data indicating that the side-chain of Met245 might be involved in one of the covalent bonds with the heme moiety. Support for attachment of the heme group to Met245 in O. fasciatus E75 LBD was obtained by the fact that both the D. melanogaster and B. mori proteins have a Thr residue at the equivalent position in helix 7 (Fig. S2).
In addition to trypsin, partial digestion of the purified O. fasciatus E75 LBD with other proteases, including pronase or proteinase K, followed by HPLC analysis also revealed the presence of hemopeptides (Fig. S8). Unfortunately, our MS/MS analysis failed to unambiguously identity additional linkages between the heme and protein moiety.
The intriguing possibility that the heme vinyl groups are involved in the covalent attachment to an amino acid side-chain was next explored using an E. coli strain defective in heme synthesis. Addition of porphyrin analogs to the bacterial medium permits the purification of heme-substituted hemoproteins that can be subsequently compared to their wild-type Fe-protoporphyrin IX counterparts (30, 40). With that in mind, we grew RP523 E. coli transformed with wild-type O. fasciatus E75 LBD in the presence of Fe(III) mesoheme, a heme derivative in which the vinyl groups have been substituted with ethyl groups. In a parallel experiment, wild-type O. fasciatus E75 LBD was also expressed in RP523 E. coli supplemented with Fe(III) protoporphyrin IX (hemin). Purification of Fe(III) protoporphyrin IX-bound O. fasciatus E75 LBD in RP523 E. coli cells revealed that the prosthetic group was successfully incorporated and, as expected, the Soret maximum was centered at 413 nm (Fig. 6A). In addition, the position of the α/β bands and the CT1 band also is undistinguishable from that shown in the O. fasciatus spectrum shown above (Fig. 1A). On the other hand, the mesoheme-saturated O. fasciatus E75 LBD revealed a sharp blue-shifted Soret maximum at 406 nm, a complete absence of CT1 band and diminished heme incorporation (approximately 69% when compared with its Fe(III) protoporphyrin IX-bound counterpart). HPLC analysis of these two purified recombinant proteins revealed that 51% of the heme in O. fasciatus E75 LBD expressed in RP523 E. coli supplemented with Fe(III) protoporphyrin IX remained covalently attached to the protein polypeptide chain (Fig. 6B, upper panel). This in is clear contrast with the O. fasciatus E75 LBD expressed in RP523 E. coli supplemented with Fe(III) mesoporphyrin IX, in which the heme group is completely unable to bind covalently to the protein moiety with 100% of the heme absorbance eluting earlier than the polypeptide chain (Fig. 6B, bottom panel). The position of this peak is consistent with the elution profile of free Fe(III) mesoporphyrin IX, that shows a sharp peak at 45 minutes (data not shown). These results strongly support the conclusion that covalent heme binding naturally occurs and clearly establishes that the vinyl groups are the site of covalent binding.
Due to the high sequence similarity at the amino acid level among the three insect E75 LBDs (Fig. S2) we concluded that very few residues present in the O. fasciatus isoform could be responsible for its different spectroscopic properties and also for the heme to protein covalent attachment. In addition, we wondered if these two phenotypes were connected. With that in mind we designed D. melanogaster/O. fasciatus and O. fasciatus/D. melanogaster chimeric E75 LBDs (Fig. S3). Since residues in helix 5 are extremely similar among the various E75 proteins, the conserved TLLKAG amino acid motif, identical in both D. melanogaster and O. fasciatus was selected as the connection part between the sequences. The absorption spectra of the Fe(III) forms of both the Dros/Onc and the Onc/Dros chimeras are shown in Fig. 7A. Intriguingly, in both cases the Soret (γ band) shows a peak at 422 nm, which is closer to the wild-type D. melanogaster E75 LBD (maximum at 424 nm) than to wild-type O. fasciatus E75 LBD (415 nm). However, the Onc/Dros chimera displayed a clear β band at 537 nm together with a small α band at 574 nm, followed by a clear CT1 band at 655 nm whereas the Dros/Onc chimera displayed a clear β band at 543 nm together with a small α band at 574 nm, followed by a faint but discernible CT1 band at 655 nm. Hence, these absorbance spectra share components from both wild-type O. fasciatus and D. melanogaster spectra although the CT1 band at 655 nm seems to arise from the interaction of the heme moiety with residues located in the N-terminal part of the O. fasciatus E75 LBD. The Fe(III) form of wild-type D. melanogaster E75 LBD displayed α/β bands at 574 and 543 nm, respectively, both being present in the chimeric constructs, albeit with a slightly blue-shifted β band. On the other hand, the Fe(III) form of wild-type O. fasciatus E75 LBD displayed a clear β band at 532 nm together with a CT1 band at 655 nm, this high spin component being clearly visible in the Onc/Dros chimera and to a lesser extent in the Dros/Onc chimera (Fig. S4). HPLC analysis of the Onc/Dros and Dros/Onc chimeras revealed that, in both cases, a significant amount of heme remained attached to the protein moiety (Fig. 7B). This suggests that a heme to protein covalent bond is retained in each of the chimeras and strongly supports the hypothesis that heme is attached to the wild-type O. fasciatus E75 LBD through at least two covalent bonds, one positioned at the N-terminus half of the polypeptide chain and the other at its C-terminal half. Quantitation of the ratio of free to bound heme in the various hemoproteins revealed that both of the chimeras partially bind heme covalently although to a lesser extent than the wild-type O. fasciatus E75 LBD (Table 2).
Spectroscopic characterization of the Met245Thr O. fasciatus E75 LBD mutant protein revealed a Soret maximum centered at 413 nm, an almost non-existing α band and a β band centered at 532 nm. Furthermore, the Met245Thr O. fasciatus E75 LBD mutant protein displayed a clear CT1 band at 655 nm (Fig. 8A), in agreement with the previous observation that the appearance of this band is the result of the interaction of the heme with residues located at the N-terminus of O. fasciatus E75 LBD. HPLC analysis of this mutant protein showed that approximately 18% of the total heme eluted as free heme, twice as much as the amount found in the wild-type protein (Table 2). Thus, elimination of Met245 did not introduce major changes in the spectroscopic properties of the hemoprotein when compared with the wild-type counterpart, although the mutation modestly increased the amount of heme that was not covalently linked to the protein moiety. Our data specifically suggest that at least two amino acid side-chains are responsible for the covalent linkage to the heme group and that the spectroscopic properties of the O. fasciatus E75 LBD are mostly determined by the proximal and distal iron ligands as well as by other amino acids present in the environment of the heme rather that by the covalent linkage to the protein backbone.
In addition to a covalent vinyl sulfonium bridge between a heme vinyl group and a Met residues such as that in myeloperoxidase, other proteins such as cytochromes P450 of the CYP4A family bind heme through the side-chain of Glu residues (20). Hence, we also created a Glu158Lys mutant, as this amino acid, present in O. fasciatus but absent in D. melanogaster and B. mori E75 LBD isoforms (Fig. S2) is predicted to be positioned at the N-terminus of helix 3 in proximity of the heme group (Fig. S7). Nonetheless, both the spectral properties and HPLC elution profile of O. fasciatus Glu158Lys mutant were virtually indistinguishable from those of its wild-type counterpart (Fig. 8B).
Next we wondered if the spectral properties and covalent linkage of the heme group to the protein were a particular characteristic of the O. fasciatus E75 LBD, or if they were shared by other insect E75 nuclear receptors. Sequence comparison showed that the homologous E75 LBD of the German cockroach (Blattella germanica) displayed a large degree of sequence identity with its O. fasciatus counterpart (Fig. S2). Thus, we cloned, expressed in bacteria and purified the E75 LBD of B. germanica (residues 148–402). As expected, the recombinant protein was purified as a hemoprotein in its ferric form with a brownish colour that highly resembled that of the O. fasciatus E75 LBD (data not shown). The Soret maximum was centered at 415 nm, the α band was almost non-existent, the β band was centered at 533 nm and a CT1 band appeared at 655 nm (Fig. 9A). These spectroscopic properties of the B. germanica E75 LBD were clearly different from those of the D. melanogaster and B. mori E75 LBDs and almost identical to those shown by the O. fasciatus E75 LBD (Fig. 1A). HPLC analysis (Fig. 9B) showed that the majority of the heme remained bound to the protein moiety in the B. germanica E75 LBD (Table 2), in agreement with the behaviour displayed by its O. fasciatus counterpart. These results led to the conclusion that the LBDs of certain E75 nuclear receptors bind the heme group covalently.
In metazoans, nuclear receptors have been classically defined as ligand activated transcription factors that allow the regulation of target genes by small lipophilic molecules such as hormones (steroids, thyroid hormones), morphogens (retinoic acid) and dietary components (fatty acids) (41, 42). Most of the nuclear receptors have distinctive modular domains designated (from the N-terminus) as A/B, C, D, E and F (42, 43). The C domain consists of the DNA-binding domain and possesses invariant Cys residues that stabilize several zinc fingers. The E domain consists of the LBD which permits nuclear receptors to function as ligand-dependent transcriptional regulators. The genome of the insect D. melanogaster encodes only 18 nuclear receptors with both a functional DNA binding domain and a LBD (44). In spite of being an extensively studied organism, many D. melanogaster nuclear receptors remain orphan, still awaiting the identification of endogenous or xenobiotic ligands. Remarkably, recent studies have shown that heme is present in the orphan nuclear receptor E75 purified from insects (3) and that heme appears associated to the E75 LBD when purified from a heterologous expression system (3, 7, 8). Soon after, two groups independently found that Rev-erbα and Rev-erbβ, the human homologs of D. melanogaster E75, also bind heme (11–13). Heme binding to these nuclear receptors can be explained assuming that they might function, perhaps even simultaneously, as sensors of cellular heme levels, redox changes or concentrations of diatomic gases, subsequently regulating the expression of a variety of genes.
Indeed, it has been suggested that D. melanogaster E75 might use heme as a ligand, hence functioning as a heme sensor (3, 8), a suggestion reinforced by the observation that heme bound to E75 at a 1:1 stoichiometry, enhanced the thermal stability of the protein and was required for proper folding and functioning of the E75 (3). Nevertheless, a caveat exists to the conclusion that heme binds and is released from E75, as a true ligand would be, since the LBD of E75 is always purified heme-saturated in bacterial expression systems (3, 7, 8) and 5 M guanidine hydrochloride is unable to release the heme moiety from purified E75 (3). We have also observed that a significant amount of heme remains bound to purified D. melanogaster E75 LBD in SDS-PAGE (data not shown). In contrast, Rev-erbα and Rev-erbβ, the mammalian homologs of insect E75 LBD, seem to be bona fide heme sensors since both proteins can be purified in the absence of heme (13, 45), isothermal titration calorimetry can be used to determine the heme binding constant for both apoRev-erbα and apoRev-erbβ (12), and atomic crystallographic data is available for both apo and heme-bound Rev-erbα (45) and Rev-erbβ (13, 46). In addition, heme binding promotes recruitment of nuclear co-repressor and histone deacetylase (12, 13, 47), with Rev-erbα recruiting HDAC3 and Rev-erbβ recruiting HDAC1 (15). Finally, whereas wild-type Rev-erbβ can repress transcription of target genes, its His602Phe mutant, which is unable to bind heme, does not display this transcriptional repressor activity in cotransfection assays (12).
Secondly, D. melanogaster E75 can detect changes in redox condition in the sense that Fe(II), but not Fe(III) E75 LBD, can bind to a peptide derived from its HR3 partner, hence suggesting that heteroassociation and repression of transcription occurs when the iron group is in its reduced form (3, 8).
Finally, the interaction of D. melanogaster E75 LBD with diatomic gases is well documented both in vivo and in vitro, although only ·NO binding to E75 LBD has been observed and characterized in vivo (3, 5–8). Binding of both CO and ·NO to the ferrous heme group of purified D. melanogaster E75 blocked its ability to bind to a peptide derived from HR3 (3). In transfection assays HR3 on its own acts as a transcriptional activator, albeit in the presence of transfected E75, its activity becomes significantly reduced. The addition of an ·NO donor, such as DETA-NO, to the cotransfected cells results in the formation of a heme-NO complex in E75, interferes with its association to HR3 and results in the rescue of its transcriptional activity (3). In vivo, ·NO release by D. melanogaster NOS in the prothoracic gland blocks the ability of E75 to interfere with HR3-mediated transcriptional activation (5). Chromatin immunoprecipitation also revealed that ·NO inhibits the recruitment of the co-repressor SMRTER by E75 (6).
In this manuscript we have characterized and compared the heme coordination of constructs of the E75 LBD hemoproteins from four different insects: Drosophila melanogaster, Bombyx mori, Oncopeltus fasciatus and Blattella germanica. Whereas B. mori E75 LBD is spectroscopically similar to its D. melanogaster counterpart, both the O. fasciatus and the B. germanica E75 LBD are purified as high-spin hemoproteins with the heme moiety covalently attached to the side-chains of at least two amino acid residues. In addition, our EPR data indicate that O. fasciatus E75 LBD is purified with a mixture of high and low spin components. Although no atomic resolution structure of insect E75 LBD is available, the crystal structure of a tryptic fragment of Rev-erbβ with heme bound indicate that Rev-erbβ Cys384 and His566 are the axial ligands for the heme group. Moreover, in light of the crystal structure of heme-bound Rev-erbβ several residues positioned in the N-terminus of helix 3 have been proposed as surrogate heme axial ligands upon conformational changes (13). It must be noted that the equivalent residues to Rev-erbβ Cys384 and His566 are conserved in D. melanogaster (Cys396, His574), B. mori (Cys193, His371), O. fasciatus (Cys144, His327) and B. germanica (Cys204, His387) E75 LBDs (Fig. S2). Nevertheless, site directed mutagenesis has revealed that in D. melanogaster Cys368, Cys385 and Cys468 are also markedly involved in heme binding (7). Sequence inspection reveals that both O. fasciatus and B. germanica E75 LBDs display an amino acid extension of the N-terminus of helix 3 that is absent from the D. melanogaster and B. mori E75 LBD isoforms. This part of the nuclear receptor is predicted to be positioned closed to the heme group and is likely to establish one of the covalent linkages to the heme moiety (Fig. S7). Our results indicate that O. fasciatus Met245 would also be covalently bound to the heme group. The absence of major spectroscopic changes in our Met245Thr mutant probably reflects the reduction of the heme vinyl group and the formation of an ethyl linkage between protein and heme moieties. This would be analogous to linkage found in cyanobacterial hemoglobins (48, 49) and in ascorbate peroxidase Ser160Met mutant (39) and different from the vinyl sulfonium linkage of myeloperoxidase (23, 38). Furthermore, strong evidence in favour of the heme vinyl groups as key elements in the covalent linkage is provided by the absence of attachment of mesoheme to the protein moiety of O. fasciatus E75 LBD when expressed in the RP523 E. coli strain. Sequence inspection reveals that very few residues present in the O. fasciatus and B. germanica isoforms could be responsible for their unique spectroscopic properties and their heme to protein covalent attachment. The results obtained using our Drosophila/Oncopeltus chimeric hemoproteins reveal that a heme attachment site must be present N-terminal and another one C-terminal of helix 3.
Our data also strongly suggest that insect E75 LBDs are very unlikely to function as heme sensors in vivo, since covalently-linked heme cannot be released from the hydrophobic pocket of the protein. In this regard, it must be noted that the tight heme binding previously reported for D. melanogaster E75 LBD and the impossibility to purify heme-free protein indicate that the binding and release of heme is also unlikely to occur in these species. It is not fully understood why heme becomes covalently bound in some proteins or even only in some members of a protein family. For example only a few members of the CYP4 family of cytochrome P450 covalently bind their heme, so covalent heme attachment satisfies a requirement that is specific for some of them. Covalent binding of the heme to the protein may allow a finer discrimination against exposure of the (-1)-methylene group to the ferryl species and thus enhance the -regiospecificity of the enzyme (20).
Our data do support the role of insect E75 LBDs as diatomic gas sensors. Notwithstanding being aerobically purified as ferric hemoproteins, in order to become regulated through ·NO and CO binding in vivo it is tempting to speculate that insect E75 LBDs should be present in their ferrous form in the cell nucleus. The Fe(II) absorbance spectra of all D. melanogaster, O. fasciatus and B. mori E75 LBDs are very similar (Fig. 2), with Soret maximum at 425 nm and sharp α/β bands at 559 and 530 nm, indicative of replacement or loss of the cysteine (thiolate) ligand, in agreement with previous data (8). Significantly, anaerobic titration revealed that the ferrous form of O. fasciatus E75 LBD readily formed both ·NO and CO complexes, implying that in vivo diatomic gases very likely regulate its transcription repression activity as well. In accordance with the proposed models (3, 7, 8), this conformational change induced by ·NO or CO binding should eventually also result in the release of O. fasciatus HR3 partner. However, the published spectra of the ·NO adducts of purified E75 and Rev-erbβ LBDs are markedly different (8). Whereas the ·NO complex of both D. melanogaster and B. mori E75 LBDs, with a maximum at 385 nm, are characteristic of a 5-coordinate hemoprotein, those of Rev-erbβ LBD (with an absorbance maximum at 418 nm) and O. fasciatus E75 LBD (with an absorbance maximum at 416 nm), correspond very likely to a 6-coordinate heme with the side-chain of a His residue trans to ·NO (3, 7, 8). Similarly, 6-coordinate Fe(II)NO complexes have been described for myoglobin (Soret at 421 nm (35)), horseradish peroxidase (Soret at 419 nm (50)), neuroglobin (Soret at 416 nm (36)) or ChCooA (Soret at 418 nm (51)).
It is unclear what advantage is provided by covalent binding of the heme in these two members, given that the heme is fully functional in the non-covalently bound state in D. melanogaster and B. mori E75 LBDs. The fact that covalently bound heme in O. fasciatus E75 LBD readily responds to the presence of CO and ·NO indicates that diatomic gases are also involved in signalling through this nuclear receptor, although its downstream effector biomolecules might be different. I nterestingly major differences have been recently discovered regarding the temporary expression pattern and the parasegment specification of the E75 gene between D. melanogaster and O. fasciatus. First of all, unlike E75A in D. melanogaster, the O. fasciatus E75A gene is a pair-rule gene required for specification of odd-numbered parasegments (26). Whereas in D. melanogaster E75A the mRNA is maternally inherited, evenly distributed and first transcribed between 6 and 8 h after egg deposition (52), the pair-rule genes, such as that of O. fasciatus E75A are expressed between 1.5 and 3 h, from the onset of cellularization through gastrulation (26). Strong evidence in favour of a different signalling pathway between D. melanogaster and its O. fasciatus and B. germanica homologs is provided by the observation that E75A expression in postembryonic development is initiated by ecdysteroid, and the appearance of the different isoforms depends on the changing ecdysteroid titer (52, 53). Conversely, whether ecdysteroids might be responsible for initiating E75 expression in O. fasciatus and B. germanica embryos is unclear. Furthermore, E75A mRNA was found in the early embryo of B. germanica, but was not correlated with an increase in ecdysteroid titer (27). Our results indicate that although all four insect E75 LBDs characterized in this work are purified as hemoproteins, the covalent heme attachment might serve different functions in vivo. The possibility therefore exists that different insect species regulate signalling through E75 nuclear receptors in divergent ways.
We are grateful to Felipe Clemente from the Proteomics unit at the Complutense University and Sara Abián for help with the HPLC. We are also indebted to Javier Merino Gracia for help with the PyMol figure. We thank Hugues Ouellet for assistance with the gas binding studies and Fatbardha Varfaj for help with the initial heme binding studies. This work was funded with Spanish grant BFU2009-10442 and NIH grants AI074824 and GM25515.