- Email: [email protected]
Identification of Cys385 in the isolated kinase insertion domain of heme-regulated eIF2α kinase (HRI) as the heme axial ligand by site-directed mutagenesis and spectral characterization
JOURNAL OF
Inorganic Biochemistry Journal of Inorganic Biochemistry 101 (2007) 1172–1179 www.elsevier.com/locate/jinorgbio
Identification of Cys385 in the isolated kinase insertion domain of heme-regulated eIF2a kinase (HRI) as the heme axial ligand by site-directed mutagenesis and spectral characterization Kyoko Hirai a, Marketa Martinkova a,1, Jotaro Igarashi a, Islam Saiful a, Seigo Yamauchi a, Samir El-Mashtoly b, Teizo Kitagawa b, Toru Shimizu a,* a b
Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan Okazaki Institute of Integrative Bioscience, National Institutes of Natural Sciences, Okazaki 444-8787, Japan Received 9 March 2007; received in revised form 29 April 2007; accepted 14 May 2007 Available online 24 May 2007
Abstract Heme-regulated eIF2a kinase (HRI) is an important enzyme that modulates protein synthesis during cellular emergency/stress conditions, such as heme deficiency in red cells. It is essential to identify the heme axial ligand(s) and/or binding sites to establish the heme regulation mechanism of HRI. Previous reports suggest that a His residue in the N-terminal region and a Cys residue in the C-terminal region trans to the His are axial ligands of the heme. Moreover, mutational analyses indicate that a residue located in the kinase insertion (KI) domain between Kinase I and Kinase II domains in the C-terminal region is an axial ligand. In the present study, we isolate the KI domain of mouse HRI and employ site-directed mutagenesis to identify the heme axial ligand. The optical absorption spectrum of the Fe(III) hemin-bound wild-type KI displays a broad Soret band at around 373 nm, while that of the Fe(II) heme-bound protein contains a band at 422 nm. Spectral titration studies conducted for both the Fe(III) hemin and Fe(II) heme complexes with KI support a 1:1 stoichiometry of heme iron to protein. Resonance Raman spectra of Fe(III) hemin-bound KI suggest that thiol is the axial ligand in a 5coordinate high-spin heme complex as a major form. Electron spin resonance (ESR) spectra of Fe(III) hemin-bound KI indicate that the axial ligands are OH and Cys. Since Cys385 is the only cysteine in KI, the residue was mutated to Ser, and its spectral characteristics were analyzed. The Soret band position, heme spectral titration behavior and ESR parameters of the Cys385Ser mutant were markedly different from those of wild-type KI. Based on these spectroscopic findings, we conclude that Cys385 is an axial ligand of isolated KI. Ó 2007 Elsevier Inc. All rights reserved. Keywords: Translation; Heme axial ligands; Kinase; ESR; Resonance Raman
1. Introduction When cells face emergency status, such as nutrition shortage, accumulation of unfolded proteins, virus infection or heme iron deficiency, they initiate protection mechanisms by terminating protein synthesis [1–5]. Protein synthesis is terminated via phosphorylation at Ser51 of *
Corresponding author. Tel.: +81 22 217 5604/5605; fax: +81 22 217 5604/5390. E-mail address: [email protected] (T. Shimizu). 1 Present address: Department of Biochemistry, Faculty of Science, Charles University, Hlavova 2030/8, 128 40 Prague 2, Czech Republic. 0162-0134/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2007.05.004
the a subunit of eukaryotic initiation factors 2 (eIF2a) by eIF2a kinases [1,2]. There are four known eIF2a kinases, specifically, GCN2, PERK, PKR and heme-regulated inhibitor (HRI), which respond to amino acid starvation, accumulation of unfolded protein (in response to ER stress), double-stranded RNA induced by virus infection, and shortage of heme iron, respectively [1–5]. In healthy red cells, the ratio of the heme iron to globin protein is 1:1. Heme iron deficiency in red cells may be caused by defects in heme synthesis or iron incorporation into protoporphyrin IX. In this case, the globin synthesis termination system is initiated to remedy the imbalance in the heme-toprotein ratio [5,6]. Under these circumstances, HRI senses
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
1173
In the present study, we expressed and purified the isolated KI domain (amino acids 241–406) of mouse HRI in Escherichia coli, which contains a putative heme binding site [9]. The heme binding characteristics and coordination structures were examined by optical absorption, resonance Raman and EPR spectroscopy. A single-site mutant, Cys385Ser, of the isolated KI domain was generated, and its spectroscopic properties compared with those of wildtype KI protein. Based on the spectral data, we conclude that the heme iron binds to Cys385 as the axial ligand, at least in the isolated KI domain.
low heme iron concentrations at the putative heme binding site, which triggers its kinase function, specifically, phosphorylation of eIF2a at Ser51, leading to the termination of protein synthesis. HRI senses the heme concentration in red cells, possibly by blocking the heme iron at the kinase active site. At normal physiological concentrations, heme binds to the kinase active site and hinders catalysis. However, at lower concentrations, heme dissociates from the active site, which is consequently exposed to the solvent, thus stimulating kinase function. Earlier studies suggest that one molecule of heme is bound to the HRI apoprotein, and that heme-bound HRI exists as a 6-coordinate low-spin complex in the presence of sufficient heme in the solution, blocking the kinase active site [7,8]. It is proposed that one of the axial ligands is His75 in the N-terminal domain, and other axial ligand trans to His75 is Cys in the C-terminal catalytic region [7,8]. However, the specific Cys residue in the C-terminal region that acts as an axial ligand remains to be identified. Rafie-Kolpin et al. [9] attempted to determine the heme axial ligands by generating two N-terminal truncated mutants and six isolated domains of rabbit HRI (Fig. 1), and examining the optical absorption spectra of the heme-bound forms. Their data show that heme binds both the N-terminal and kinase insertion (KI) domains of HRI, the latter located between the Kinase I and Kinase II domains (Fig. 1). However, to our knowledge, no detailed heme binding or other spectroscopic data on the isolated wild-type or Cys mutant KI proteins have been reported to date. Further comprehensive analyses are required to identify the Cys residue in the KI domain that acts as the axial ligand to the heme iron.
2. Materials and methods 2.1. Materials Restriction and modification enzymes were acquired from Takara Bio (Otsu, Japan), Toyobo (Osaka, Japan), New England Biolabs (Beverly, MA, USA) and Nippon Roche (Tokyo, Japan). Other chemicals were purchased from Wako Pure Chemicals (Osaka, Japan). 2.2. Plasmid construction Wild-type HRI comprising residues 1–619 from mouse liver cDNA originally cloned in pET-28a [7] was used as a PCR template. The isolated KI domain (amino acids 241–406) was amplified by PCR with a 5 0 -sense primer containing BamHI and NdeI sites (5 0 -CGGATCCCATATGCAGCCACAAGACAGAGTTCCG-3 0 ) and a 3 0 -antisense primer containing a stop codon and EcoRI site (5 0 GGAATTCTTACTCCCGGCTCCGCTTGTTGTTCC-3 0 ).
Cys
KD
Fe NTD His
Full-length wild type
NTD NTD NTD
(1-619)
KD
(145-619)
(1-144)
Kinase I
Characterize this site
KD
(145-240) Kinase Insertion
(241-406) Kinase II
(407-540) CTD
(541-619) 0
Fig. 1. The mouse HRI protein (619 amino acids) is composed of several domains, including N-terminal, Kinase I, kinase insertion, Kinase II, and Cterminal domains [9]. Here, we examine the spectral characteristics of the isolated kinase insertion domain (residues 241–406). The KI domain in Ref. [9] has residues 232–420.
1174
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
Amplified products were initially subcloned into the pCRä4-TOPO vector (Invitrogen, Carlsbad, CA) to confirm the sequences. The correct products were digested with BamHI and EcoRI, and subcloned into the pGEX-6P2 vector. Single-site mutants, Cys385Ser and Cys385Ala, were constructed with the QuikChangeä mutagenesis kit (Stratagene, La Jolla, CA) using the following primers: 5 0 -sense (5 0 -CAGATGCAGCTGAGCGAGCTCTCCCTGTGG-3 0 ) for Cys385Ser and (5 0 -CAGATGCAGCTGGCGGAGCTCTCCCTGTGG-3 0 ) for Cys385Ala; 3 0 -antisense (5 0 -CCACAGGGAGAGCTCGCTCAGCTGCATCTG-3 0 ) for Cys385Ser and (5 0 -CCACAGGGAGAGCTCCGCCAGCTGCATCTG-3 0 ) for Cys385Ala. 2.3. Protein expression and purification Glutathione-S-transferase-tagged (GST-tagged) KI was expressed in E. coli BL21 (DE3) Codon Plus RIL (Stratagene) harboring pGEX-6P2/KI, and purified as described previously [7,8,12] with some modifications. Briefly, cell lysates containing heme-free GST-tagged KI were filtered with a 0.45 mm filter (Millipore Corporation, Bedford, MA), and subjected to Glutathione Sepharose 4B chromatography (GE Healthcare, Bucks, UK). Solutions eluted with buffer containing reduced glutathione were collected and applied to Sephadex G-25 (HiTrap) (GE Healthcare) using the AKTA system (GE Healthcare) to remove reduced glutathione in buffer. The GST tag was digested with PreScissionä protease, and purified by Sephadex G-25 (HiTrap) and GS-Trap (GE Healthcare) column chromatography with the AKTA system to remove uncleaved KI and PreScission protease. Purified GST-free KI proteins were >85% homogenous, as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (15% gel), followed by staining with Coomassie Brilliant Blue R250. Protein concentrations were determined using the Coomassie Brilliant Blue dye binding method (Nacalai Tesque, Kyoto, Japan). The final protein yields were 0.18, 0.25 and 0.07 mg/l of culture for wild-type, Cys385Ser and Cys 385Ala KI, respectively. The expression efficiency of the Cys385Ala mutant was very low. Most of the protein was retained within inclusion bodies, and thus difficult to solubilize. Therefore, we only examined the spectroscopic characteristics of the wild-type and Cys385Ser mutant KI proteins.
2.5. Resonance Raman spectra The KI domain of HRI (40 lM in 50 mM Tris–HCl, pH 8.0) was placed in an airtight spinning cell with a rubber septum. Raman scattering was excited at 406.7 nm with a Kr ion laser (BeamLok 2060, Spectra-Physics, Mountain View, CA) and 363.8 nm with an Ar ion laser (Spectra Physics, BeamLok 2080). The excitation light was focused into the cell at a laser power of 5 mW for the Fe(III) complex. Raman spectra were detected with a liquid N2-cooled CCD detector (Rober Scientific, Spec-10:400B/LN, Trenton, NJ) attached to a SPEX750M single polychromator (Jobin Yvon, Longjumeau, France). Raman shifts were calibrated with indene, acetone, CCl4, and an aqueous solution of ferrocyanide. 2.6. ESR spectra The Fe(III) hemin-bound KI (200 lM) complex was prepared in 50 mM Tris–HCl, pH 8.0, containing 50% glycerol at 25 °C. ESR spectra were recorded on a JEOL FE3X spectrometer (Tokyo, Japan) at 30 K. The magnetic field was calibrated using an NMR gauss meter (Echo Electronics, Hadsund, Denmark; model EFM-2000), and the temperature controlled with an Oxford 900 cryosystem [7,8]. 3. Results and discussion 3.1. Protein expression and purification At the beginning of this study, we attempted to express N-terminal His-tagged KI. Unfortunately, we could not obtain sufficient amounts of the soluble protein, in contrast to the report by Rafie-Kolpin et al. [9]. Therefore, the GST tag was employed to enhance protein solubility and avoid any interactions between His tag and heme. PreScission protease was applied to remove the GST tag. Fe(III) heme reconstitution was performed to ensure a 1:1 stoichiometry of KI protein:heme after purification. The expression efficiency and solubility of the Cys385Ala mutant were very low. Consequently, we only examined the spectroscopic characteristics of wild-type and Cys385Ser mutant KI proteins.
2.4. Optical absorption spectra
3.2. Optical absorption spectra of the KI domain
Optical absorption spectra were collected using Shimadzu UV-2500 and Shimadzu Multi Spec 1500 spectrophotometers (Kyoto, Japan) maintained at 25 °C. Spectral changes were recorded under both aerobic and anaerobic conditions in 50 mM Tris–HCl, pH 8.0, at 25 °C. The reaction mixture was incubated for 10 min prior to spectroscopic measurements to ensure stable solution temperature and proper coordination between Fe(III)–hemin or Fe(II)– heme and the protein. Experiments were performed at least three times for each complex.
We examined the optical absorption spectra of the heme iron-bound KI domain. The Soret band of wild-type KI with Fe(III) hemin was located at around 373 nm with a small shoulder at 412 nm (Fig. 2a; Table 1). The bandwidth was broader than that of full-length wild-type HRI enzyme. In view of these results, we suggest that the spectrum contains at least two species. The same peak around 370 nm was observed for the heme complexes of iron regulatory protein 2, another heme sensor protein [15], suggesting that the band around 370 nm is not due to a
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
Fe(III) Fe(II) Fe(II)CO Fe(II)NO
Absorbance
0.3 389 373
0.2
420 422
527
539 559
536 539 569
0.1 0
b 0.3 Absorbance
a
Fe(III) Fe(II) Fe(II)CO Fe(II)NO
389 420 391
0.2
1175
535 558 614
420
0.1
538
564
0.0
300
400
500
600
700
300
Wavelength (nm)
400
500
600
700
Wavelength (nm)
Fig. 2. Optical absorption spectra of heme-bound isolated wild-type (a) and Cys385Ser mutant (b) KI complexed to Fe(III) (bold solid line), Fe(II) (thin solid line in gray), Fe(II)–CO (dotted line) and Fe(II)–NO (broken line in red). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 1 Optical absorption spectra (nm) of the heme-bound full-length and isolated kinase insertion (KI) domain HRI complexes Fe(III) hemin
Wild-type full-length Wild-type KI Cys385Ser KI a
Fe(II) heme
Fe(II) heme–CO
Fe(II) heme–NO
Soret
Visible
Soret
Visible
Soret
Visible
Soret
Visible
366, 421 373, 412a 391
544 539 614
426 422 420
532, 559 527, 559 558
421 420 420
538, 564 536, 569 538, 564
398 389 389
538 539 535
Shoulder.
non-specific binding of the Fe(III)–hemin to the KI domain. Also, note that the spectrum of the free Fe(III)– hemin complex has broad peaks at approximately 387 and 357 nm, corroborating that the 373 nm species is not the free Fe(III)–hemin complex. Upon reduction of the heme iron with sodium dithionite, a narrow Soret band ascribed to the Fe(II) heme-bound wild-type KI domain was observed at 422 nm (Fig. 2a; Table 1). The absorption spectrum of the Fe(II) heme–CO complex of wild-type KI contained a Soret peak at 420 nm, similar to full-length wild-type HRI enzyme. The absence of the 450 nm peak upon CO binding to the Fe(II) heme–KI proteins implies that the Cys residue is not coordinated to the Fe(II) heme iron, being different from cytochrome P450 proteins. Our spectral findings differ from those of Rafie-Kolpin et al. [9] who report Soret absorption at 415 nm for the Fe(III) KI domain. The differences between the two sets of spectral results may be attributed to the His-tagged protein used in their experiments, in place of our GST-tag free protein. Since Cys385 is the only cysteine residue in the KI domain, we generated a Cys385Ser KI mutant, and examined the corresponding optical absorption spectra. The absorption spectra of Fe(III) hemin- and Fe(II) hemebound Cys385Ser KI were distinct from those of the wild-type KI domain in that the Soret bands of the mutant proteins were located at 391 and 420 nm, respectively (Fig. 2b; Table 1). The data suggest that ligand switching occurs in the mutant protein. In contrast, the spectrum of the Fe(II)–CO-bound Cys385Ser KI domain was similar to that of wild-type KI. Since there are four His residues in the KI domain, it is possible that ligand exchange occurs as a result of heme reduction and/or mutation of Cys385.
The full-length wild-type HRI protein is a 5-coordinated Fe(II) heme–NO complex displaying Soret absorption maximum at 398 nm [7]. In our experiments, the Fe(II) heme–NO complex of the isolated wild-type KI domain displayed a similar Soret band at 389 nm, supporting the presence of the 5-coordinated Fe(II)–NO complex. Interestingly, the Cys385Ser KI domain exists as a 5-coordinated Fe(II)–NO complex with a Soret absorption peak at 389 nm, similar to the cystathionine b-synthase complex [14]. Thus, the surrounding environment of the heme binding site of the mutant appears to retain a favorable conformation for the 5-coordinated Fe(II)–NO complex, despite subtle structural changes as a result of the mutation. 3.3. Heme binding properties of the KI domain We examined the heme binding characteristics of the Fe(III) hemin and Fe(II) heme complexes of wild-type and Cys385Ser mutant KI proteins. Differences in the absorption spectral changes of wild-type KI domain caused by the addition of Fe(III) hemin disclose a clear change in linearity, and support a heme to protein stoichiometry of 1:1 (Fig. 3a and c). The difference spectral titration of Fe(III) hemin with the wild-type KI domain also displayed a distinct positive peak at 360 nm (Fig. 3a). A Soret peak at around 360 nm characteristic of the thiol-coordinated high-spin complex of Fe(III) hemin-bound protein was observed for Fe(III) hemin-bound N-terminal truncated HRI protein [8,15]. Our findings suggest that the Cys residue in wild-type KI is an axial ligand for the Fe(III) hemin-bound form. Absolute optical absorption spectral changes of Fe(II) heme caused by addition of the wild-type KI domain
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
ΔAbsorbance
a 0.5
c
0.4 0.3 0.2 0.1 0.0 300
400
500
600
700
ΔAbsorbance at 360 nm
1176
0.5 0.4 0.3 0.2 0.1 0.0 0.0
Wavelength (nm)
d
Absorbance
1.0 0.8 0.6 0.4 0.2 0.0 400
500
600
700
Absorbance at 426 nm
b
300
0.5
1.0
1.5
2.0
Fe(III) hemin / apo KI
0.8 0.6 0.4 0.2 0.0 0.0
0.5
1.0
1.5
2.0
Fe(ll) heme/ apo KI
Wavelength (nm)
Fig. 3. Differences in absorption spectral changes of wild-type KI (10 lM) caused by addition of Fe(III) hemin (1 lM each) (a) and absolute optical absorption spectral changes of wild-type KI (10 lM) caused by addition of Fe(II) heme (1 lM each) (b). Spectral titration plots (filled circles) monitored at 360 nm and 426 nm for Fe(III) hemin (c) and Fe(II) heme (d), respectively, reveal a 1:1 stoichiometry of heme:protein for both complexes. The open circles in (d) represent the increase in absorbance of the free Fe(II) heme.
the possibility that axial ligand exchange occurs from Cys to another residue is suggested. Titration experiments were conducted for the Cys385Ser KI mutant. We did not observe any distinct changes in linearity reflecting a 1:1 stoichiometry of Fe(III) hemin binding to the mutant protein (Fig. 4a and c). The peak at 360 nm characteristic of the thiol-coordinated high-spin complex of Fe(III) hemin-bound mutant protein was
Δ
Δ
additionally support a 1:1 stoichiometry of heme to protein, although the change in linearity is less clear, compared with titration of the Fe(III) hemin complex (Fig. 3b and d). We believe that Fe(II) heme binding is relatively weak, compared to that of Fe(III) hemin, particularly when the axial ligand is thiol. This is a reasonable assumption, since the thiol negative charge and relatively electron-sufficient Fe(II) (compared with Fe(III)) repel each other. Rather,
Fig. 4. Differences in absorption spectral changes of Cys385Ser KI (10 lM) caused by addition of Fe(III) hemin (1 lM each) (a) and absolute optical absorption spectral changes of Cys385Ser KI (10 lM) caused by addition of Fe(II) heme (1 lM each) (b). Spectral titration plots (filled circles) monitored at 360 nm and 421 nm for Fe(III) hemin (c) and Fe(II) heme (d), respectively, disclose that the stoichiometry of heme:protein is not 1:1 for both mutant complexes. Specifically, for Fe(III) hemin binding, no alterations in the linearity of absorption change are observed, whereas for Fe(II) heme binding, the stoichiometry of heme:protein is 1:2. The open circles in (d) represent the increase in absorbance of the free Fe(II) heme.
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
significantly lower than that of wild-type protein (Figs. 3a and 4a). The data collectively imply that Cys385 is an axial ligand of the Fe(III) hemin-bound wild-type KI domain. For Fe(II) heme binding to Cys385Ser KI, the heme to protein ratio was calculated as 1:2, based on spectral changes of the mutant protein. We assume that unusual or non-specific binding occurs in this case. Therefore, titration of the Fe(II) heme complex into the Cys385Ser mutant led to ambiguous results, confirming that Cys385 is the axial ligand for Fe(II) heme-bound wild-type KI (Fig. 4b and d). In total, 4 His and 2 Met residues exist in the KI domain that may bind to heme as axial ligands.
c). Accordingly, we propose that the Fe(III) hemin-bound KI domain contains a Fe–Cys bond, similar to the fulllength HRI protein. The resonance Raman spectra indicate that the Fe(III) hemin-bound KI domain is in equilibrium between the 5-coordinated Fe(III)–Cys and 6-coordinated water or OH–Fe(III)–Cys complexes. In general, the 6coordinate low-spin heme does not show Fe–Cys or Fe– His stretching modes. 3.5. Electron spin resonance (ESR) spectra of the Fe(III) hemin-bound KI domain
406.7 nm 363.8 nm
KI
Intensity (a.u.)
377
341
415
338 ν (Fe-Cys)
Intensity (a.u.)
1620 1627
1502
1553 1582
1490
1428 1442
λ ex = 406.7nm
WT
ν8
c
HRI, pH 8.0
1374
Intensity (a.u.)
WT
HRI, pH 8.0 λex = 363.8 nm
675
b
347
Resonance Raman spectra provide valuable information about the heme coordination structure [10–12]. The marker band, m3, at the high frequency region of the KI domain is located at 1490 cm1 with a small band at 1502 cm1 (Fig. 5a), implying that the 5-coordinate high-spin complex is the major species, with 6-coordinate low-spin complex as a minor species, consistent with absorption spectral findings. This finding is distinct from that for the full-length HRI protein where the band at 1502 cm1 ascribed to the 6-coordinated low-spin complex is the major species, with a minor band at 1490 cm1 attributed to the 5-coordinated high-spin complex (Fig. 5a). In the low frequency region, the band ascribed to m(Fe– Cys) was observed at 338 cm1 (with laser excitation at 363.8 nm) for the Fe(III) hemin-bound full-length wildtype HRI protein (upper spectra in Fig. 5b and c). Similar bands in this region were observed for the Fe(III) heminbound KI domain protein (lower spectra in Fig. 5b and
338 ν (Fe-Cys)
Crystal field analysis of the ESR spectra of Fe(III) hemin low-spin signals reveals tetragonality and rhombicity [7,8]. These ESR parameters are useful to identify the axial ligands of low-spin Fe(III) hemin complexes. Thus, axial ligands of the Fe(III) hemin-bound full-length wildtype HRI were identified as His/Cys, whereas those of the N-terminal truncated HRI protein were OH/Cys (Fig. 6c, right) [7,8]. Fe(III) hemin-bound wild-type KI contained a mixture of high-spin and low-spin complexes, and the low-spin complex gave g values of 2.43, 2.26 and 1.91 (Fig. 6a). ESR parameters of the low-spin complex revealed that OH/Cys are the axial ligands of the Fe(III) hemin-bound wild-type KI domain, similar to the N-terminal truncated HRI protein (Fig. 6c, right). ESR spectra of the Fe(III) hemin-bound Cys385Ser mutant of the KI domain contained a mixture of several species (Fig. 6b). Since the ESR parameters of the low-spin complex are located outside the region analyzed, it was not feasible to identify the axial ligands of the Fe(III) hemin-bound Cys385Ser mutant KI domain. Therefore, the coordination of Cys385 in the Fe(III) hemin-bound wild-type KI was further corroborated from the ESR spectra.
3.4. Resonance Raman spectra of the KI domain
a
1177
WT
KI
KI
1400
1500
1600
Raman Shift (cm-1)
1700
300
400
500
600
Raman Shift (cm-1)
700
320
340
360
Raman Shift (cm-1)
Fig. 5. High frequency (a) and low frequency (b) regions of resonance Raman spectra of Fe(III) heme-bound wild-type full-length HRI (upper spectra) and KI (lower spectra) proteins. Spectra of the low frequency region are expanded in (c). For both proteins, bands at 338–341 cm1 (when excited at 363.8 nm) are assigned to the Fe–S stretching band, while close to it m8 band is observed at 347 cm1 upon excitation at 406.7 nm. This observation suggests that one of the axial ligands is Cys for the KI domain protein, similar to that of full-length HRI. Note that only the 5-coordinated Fe(III)–Cys complex, but not the 6-coordinated low-spin heme, shows the Fe–Cys stretching mode. WT, full-length wild-type HRI; KI, isolated wild-type KI domain.
OH–
1.99 1.91
6.09
a
2.43 2.26
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
4.35
1178
Fe3+
Fe3+
S
S
Cys385
Cys385
373 nm High spin
~ 412 nm Low spin e-
100
200
300
400
Magnetic Field (mT)
O C
100
200
2.00 1.91
2.28
2.46
6.08
b
4.39
Fe2+
300
400
Magnetic Field (mT)
Rhombicity |R / μ|
c
0.60
CO
X' Fe2+
X
X
420 nm Low spin
422 nm Low spin
Fig. 7. Proposed coordination structures of the isolated KI domain of HRI. The thiol group of Cys385 would be bound the Fe(III) hemin complex, at least in the isolated KI domain. The Fe(III) hemin-bound isolated KI domain would be in equilibrium between the 5-coordinated high-spin Fe(III)–Cys (major form) and 6-coordinated low-spin water or OH–Fe(III)–Cys (minor form) complexes based on the resonance Raman and ESR spectra. The axial ligand, Cys385, of the Fe(III) complex would switch to another unidentified amino acid residue upon reduction in that the axial ligand at the proximal side for both the Fe(II) and Fe(II)–CO complexes is no more thiol of Cys.
His/His
0.55
KI WT
0.50
His/Cys
0.45
-
HO /Cys
0.40 0.35 3
4
5
6
Tetragonality |μ / λ| Fig. 6. ESR spectra (left) of the wild-type (a) and Cys385Ser (b) proteins of the isolated KI domain of HRI. Based on the crystal field diagram for Fe(III) low-spin complexes (c), it is suggested that the isolated wild-type KI protein has OH/Cys as axial ligands, in contrast to the full-length wild-type HRI enzyme where His/Cys are possibly the axial ligands. The areas surrounded by the solid and dashed rectangles and dotted circles contain His/Cys, His/His, and Cys/water or hydroxyl anion as the axial ligands, respectively. The axial ligands of Cys385Ser KI could not be identified, since the parameters are located outside the region examined (data not shown). ESR spectra at the high spin region of the both wildtype and Cys385Ser mutant proteins did not provide valuable information about their coordination structures. WT, full-length wild-type HRI; KI, the isolated wild-type KI domain.
4. Summary Based on the optical, resonance Raman and EPR spectral results of the wild-type and Cys385Ser mutant proteins of the isolated KI domain, we conclude that the heme iron binds to Cys385 as an axial ligand for the Fe(III) hemin
complex, at least in the isolated KI domain. Perhaps the Fe(III) hemin complex would be in the equilibrium between two species with the high- and low-spin states (Fig. 7). However, upon reduction of the heme iron, thiol of Cys385 would dissociated from the heme iron and unidentified amino acid residue(s) would then become the axial ligands for both the Fe(II) heme and Fe(II) heme– CO complexes (Fig. 7). Fig. 8 summarizes the proposed structures of the HRI proteins. In the full-length wild-type protein (WT), His in the N-terminal region and Cys from the Kinase Domain at the C-terminal region are the candidate axial ligands of the Fe(III) heme. In the N-terminal truncated mutant (D145), OH and Cys in the Kinase Domain appear to be the axial ligands of the Fe(III) heme for the 6-coordinated low-spin complexes. However, the major form would be probably the 5-coordinated high-spin complexes with only one axial ligand of Cys385 in the Fe(III) hemin complex of the isolated KI domain. On the other hand, we cannot exclude the possibility that isolation of the KI domain from full-length wild-type HRI protein causes global protein conformational changes. Conformational changes in the N-terminal domain were observed upon its isolation so that heme ligand switching occurs upon the isolation of the domain. Specifically, His/His were proposed as the axial ligands of the isolated N-terminal domain of HRI (Fig. 7, far right), in contrast to His/ Cys for the full-length HRI protein (Fig. 7, far left) [7,12,13]. Moreover, full-length HRI was converted from the hexameric to the trimeric state upon removal of the
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
Cys
KD
Fe NTD NTD His
WT His/Cys
Cys KD Fe OH-
Δ145 —HO/Cys
KD Cys KI Fe Fe
1179 His Fe
NTD His
KI Cys
NTD His/His
Fig. 8. Proposed structures of HRI proteins. WT, full-length wild-type HRI protein; D145, N-terminal truncated mutant of WT [9]; KI, isolated KI domain; NTD, isolated N-terminal domain [12,13].
N-terminal domain [8]. Examination of the isolated wildtype KI domain revealed a higher oligomeric state than hexamer. The regions surrounding the KI domain may thus be important to maintain its distinct structure.
Dr. H. Kurokawa for help with the initial stages of the work and Dr. A. Tanaka for valuable discussions.
5. Abbreviations
[1] A.G. Hinnebusch, Nat. Struct. Mol. Biol. 12 (2005) 835–838. [2] R.C. Wek, H.Y. Jiang, T.G. Anthony, Biochem. Soc. Trans., Part 1 34 (2006) 7–11. [3] D. Ron, Science 313 (2006) 52–53. [4] J. Marx, Science 313 (2006) 1564–1566. [5] T.E. Dever, Cell 108 (2002) 545–556. [6] J.J. Chen, in: N. Soneberg, J.W.B. Hershey, M.B. Mathews (Eds.), Translational Control of Gene Expression, Cold Spring Harbor Laboratory, Cold spring Harbor, NY, 2000, pp. 529–546. [7] J. Igarashi, A. Sato, T. Kitagawa, T. Yoshimura, S. Yamauchi, I. Sagami, T. Shimizu, J. Biol. Chem. 279 (2004) 15752–15762. [8] M. Miksanova, J. Igarashi, M. Minami, I. Sagami, S. Yamauchi, H. Kurokawa, T. Shimizu, Biochemistry 45 (2006) 9894–9905. [9] M. Rafie-Kolpin, P.J. Chefalo, Z. Hussain, J. Hahn, S. Uma, R.L. Matts, J.J. Chen, J. Biol. Chem. 275 (2000) 5171–5178. [10] T. Uchida, E. Sato, A. Sato, I. Sagami, T. Shimizu, T. Kitagawa, J. Biol. Chem. 280 (2006) 21358–21368. [11] T. Tosha, N. Kagawa, T. Ohta, S. Yoshioka, M.R. Waterman, T. Kitagawa, Biochemistry 45 (2006) 5631–5640. [12] J. Igarashi, A. Sato, T. Kitagawa, I. Sagami, T. Shimizu, Biochim. Biophys. Acta 1650 (2003) 99–104. [13] H. Ishikawa, B.G. Yun, S. Takahashi, H. Hori, R.L. Matts, K. Ishimori, I. Morishima, J. Am. Chem. Soc. 124 (2002) 13696–13697. [14] S. Taoka, R. Banerjee, J. Inorg. Biochem. 87 (2001) 245–251. [15] H. Ishikawa, M. Kato, H. Hori, K. Ishimori, T. Kirisako, F. Tokunaga, K. Iwai, Mol. Cell 19 (2005) 171–181.
HRI
heme-regulated inhibitor or heme-regulated eIF2a kinase eIF2a the a subunit of eukaryotic initiation factor 2 WT full-length wild-type HRI KI domain kinase insertion domain of HRI (amino acids 241–406 for mouse) C385S the Cys385Ser mutant of the isolated KI domain Fe(III) hemin Fe(III)–protoporphyrin IX complex Fe(II) heme Fe(II)–protoporphyrin IX complex GST-tag glutathione-S-transferase-tag
Acknowledgements This work was supported, in part, by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to M.M. and T.S.) and a Postdoctoral Fellowship for Foreign Researchers to M.M. from the Japan Society for the Promotion of Science. We thank
References
Inorganic Biochemistry Journal of Inorganic Biochemistry 101 (2007) 1172–1179 www.elsevier.com/locate/jinorgbio
Identification of Cys385 in the isolated kinase insertion domain of heme-regulated eIF2a kinase (HRI) as the heme axial ligand by site-directed mutagenesis and spectral characterization Kyoko Hirai a, Marketa Martinkova a,1, Jotaro Igarashi a, Islam Saiful a, Seigo Yamauchi a, Samir El-Mashtoly b, Teizo Kitagawa b, Toru Shimizu a,* a b
Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan Okazaki Institute of Integrative Bioscience, National Institutes of Natural Sciences, Okazaki 444-8787, Japan Received 9 March 2007; received in revised form 29 April 2007; accepted 14 May 2007 Available online 24 May 2007
Abstract Heme-regulated eIF2a kinase (HRI) is an important enzyme that modulates protein synthesis during cellular emergency/stress conditions, such as heme deficiency in red cells. It is essential to identify the heme axial ligand(s) and/or binding sites to establish the heme regulation mechanism of HRI. Previous reports suggest that a His residue in the N-terminal region and a Cys residue in the C-terminal region trans to the His are axial ligands of the heme. Moreover, mutational analyses indicate that a residue located in the kinase insertion (KI) domain between Kinase I and Kinase II domains in the C-terminal region is an axial ligand. In the present study, we isolate the KI domain of mouse HRI and employ site-directed mutagenesis to identify the heme axial ligand. The optical absorption spectrum of the Fe(III) hemin-bound wild-type KI displays a broad Soret band at around 373 nm, while that of the Fe(II) heme-bound protein contains a band at 422 nm. Spectral titration studies conducted for both the Fe(III) hemin and Fe(II) heme complexes with KI support a 1:1 stoichiometry of heme iron to protein. Resonance Raman spectra of Fe(III) hemin-bound KI suggest that thiol is the axial ligand in a 5coordinate high-spin heme complex as a major form. Electron spin resonance (ESR) spectra of Fe(III) hemin-bound KI indicate that the axial ligands are OH and Cys. Since Cys385 is the only cysteine in KI, the residue was mutated to Ser, and its spectral characteristics were analyzed. The Soret band position, heme spectral titration behavior and ESR parameters of the Cys385Ser mutant were markedly different from those of wild-type KI. Based on these spectroscopic findings, we conclude that Cys385 is an axial ligand of isolated KI. Ó 2007 Elsevier Inc. All rights reserved. Keywords: Translation; Heme axial ligands; Kinase; ESR; Resonance Raman
1. Introduction When cells face emergency status, such as nutrition shortage, accumulation of unfolded proteins, virus infection or heme iron deficiency, they initiate protection mechanisms by terminating protein synthesis [1–5]. Protein synthesis is terminated via phosphorylation at Ser51 of *
Corresponding author. Tel.: +81 22 217 5604/5605; fax: +81 22 217 5604/5390. E-mail address: [email protected] (T. Shimizu). 1 Present address: Department of Biochemistry, Faculty of Science, Charles University, Hlavova 2030/8, 128 40 Prague 2, Czech Republic. 0162-0134/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2007.05.004
the a subunit of eukaryotic initiation factors 2 (eIF2a) by eIF2a kinases [1,2]. There are four known eIF2a kinases, specifically, GCN2, PERK, PKR and heme-regulated inhibitor (HRI), which respond to amino acid starvation, accumulation of unfolded protein (in response to ER stress), double-stranded RNA induced by virus infection, and shortage of heme iron, respectively [1–5]. In healthy red cells, the ratio of the heme iron to globin protein is 1:1. Heme iron deficiency in red cells may be caused by defects in heme synthesis or iron incorporation into protoporphyrin IX. In this case, the globin synthesis termination system is initiated to remedy the imbalance in the heme-toprotein ratio [5,6]. Under these circumstances, HRI senses
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
1173
In the present study, we expressed and purified the isolated KI domain (amino acids 241–406) of mouse HRI in Escherichia coli, which contains a putative heme binding site [9]. The heme binding characteristics and coordination structures were examined by optical absorption, resonance Raman and EPR spectroscopy. A single-site mutant, Cys385Ser, of the isolated KI domain was generated, and its spectroscopic properties compared with those of wildtype KI protein. Based on the spectral data, we conclude that the heme iron binds to Cys385 as the axial ligand, at least in the isolated KI domain.
low heme iron concentrations at the putative heme binding site, which triggers its kinase function, specifically, phosphorylation of eIF2a at Ser51, leading to the termination of protein synthesis. HRI senses the heme concentration in red cells, possibly by blocking the heme iron at the kinase active site. At normal physiological concentrations, heme binds to the kinase active site and hinders catalysis. However, at lower concentrations, heme dissociates from the active site, which is consequently exposed to the solvent, thus stimulating kinase function. Earlier studies suggest that one molecule of heme is bound to the HRI apoprotein, and that heme-bound HRI exists as a 6-coordinate low-spin complex in the presence of sufficient heme in the solution, blocking the kinase active site [7,8]. It is proposed that one of the axial ligands is His75 in the N-terminal domain, and other axial ligand trans to His75 is Cys in the C-terminal catalytic region [7,8]. However, the specific Cys residue in the C-terminal region that acts as an axial ligand remains to be identified. Rafie-Kolpin et al. [9] attempted to determine the heme axial ligands by generating two N-terminal truncated mutants and six isolated domains of rabbit HRI (Fig. 1), and examining the optical absorption spectra of the heme-bound forms. Their data show that heme binds both the N-terminal and kinase insertion (KI) domains of HRI, the latter located between the Kinase I and Kinase II domains (Fig. 1). However, to our knowledge, no detailed heme binding or other spectroscopic data on the isolated wild-type or Cys mutant KI proteins have been reported to date. Further comprehensive analyses are required to identify the Cys residue in the KI domain that acts as the axial ligand to the heme iron.
2. Materials and methods 2.1. Materials Restriction and modification enzymes were acquired from Takara Bio (Otsu, Japan), Toyobo (Osaka, Japan), New England Biolabs (Beverly, MA, USA) and Nippon Roche (Tokyo, Japan). Other chemicals were purchased from Wako Pure Chemicals (Osaka, Japan). 2.2. Plasmid construction Wild-type HRI comprising residues 1–619 from mouse liver cDNA originally cloned in pET-28a [7] was used as a PCR template. The isolated KI domain (amino acids 241–406) was amplified by PCR with a 5 0 -sense primer containing BamHI and NdeI sites (5 0 -CGGATCCCATATGCAGCCACAAGACAGAGTTCCG-3 0 ) and a 3 0 -antisense primer containing a stop codon and EcoRI site (5 0 GGAATTCTTACTCCCGGCTCCGCTTGTTGTTCC-3 0 ).
Cys
KD
Fe NTD His
Full-length wild type
NTD NTD NTD
(1-619)
KD
(145-619)
(1-144)
Kinase I
Characterize this site
KD
(145-240) Kinase Insertion
(241-406) Kinase II
(407-540) CTD
(541-619) 0
Fig. 1. The mouse HRI protein (619 amino acids) is composed of several domains, including N-terminal, Kinase I, kinase insertion, Kinase II, and Cterminal domains [9]. Here, we examine the spectral characteristics of the isolated kinase insertion domain (residues 241–406). The KI domain in Ref. [9] has residues 232–420.
1174
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
Amplified products were initially subcloned into the pCRä4-TOPO vector (Invitrogen, Carlsbad, CA) to confirm the sequences. The correct products were digested with BamHI and EcoRI, and subcloned into the pGEX-6P2 vector. Single-site mutants, Cys385Ser and Cys385Ala, were constructed with the QuikChangeä mutagenesis kit (Stratagene, La Jolla, CA) using the following primers: 5 0 -sense (5 0 -CAGATGCAGCTGAGCGAGCTCTCCCTGTGG-3 0 ) for Cys385Ser and (5 0 -CAGATGCAGCTGGCGGAGCTCTCCCTGTGG-3 0 ) for Cys385Ala; 3 0 -antisense (5 0 -CCACAGGGAGAGCTCGCTCAGCTGCATCTG-3 0 ) for Cys385Ser and (5 0 -CCACAGGGAGAGCTCCGCCAGCTGCATCTG-3 0 ) for Cys385Ala. 2.3. Protein expression and purification Glutathione-S-transferase-tagged (GST-tagged) KI was expressed in E. coli BL21 (DE3) Codon Plus RIL (Stratagene) harboring pGEX-6P2/KI, and purified as described previously [7,8,12] with some modifications. Briefly, cell lysates containing heme-free GST-tagged KI were filtered with a 0.45 mm filter (Millipore Corporation, Bedford, MA), and subjected to Glutathione Sepharose 4B chromatography (GE Healthcare, Bucks, UK). Solutions eluted with buffer containing reduced glutathione were collected and applied to Sephadex G-25 (HiTrap) (GE Healthcare) using the AKTA system (GE Healthcare) to remove reduced glutathione in buffer. The GST tag was digested with PreScissionä protease, and purified by Sephadex G-25 (HiTrap) and GS-Trap (GE Healthcare) column chromatography with the AKTA system to remove uncleaved KI and PreScission protease. Purified GST-free KI proteins were >85% homogenous, as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (15% gel), followed by staining with Coomassie Brilliant Blue R250. Protein concentrations were determined using the Coomassie Brilliant Blue dye binding method (Nacalai Tesque, Kyoto, Japan). The final protein yields were 0.18, 0.25 and 0.07 mg/l of culture for wild-type, Cys385Ser and Cys 385Ala KI, respectively. The expression efficiency of the Cys385Ala mutant was very low. Most of the protein was retained within inclusion bodies, and thus difficult to solubilize. Therefore, we only examined the spectroscopic characteristics of the wild-type and Cys385Ser mutant KI proteins.
2.5. Resonance Raman spectra The KI domain of HRI (40 lM in 50 mM Tris–HCl, pH 8.0) was placed in an airtight spinning cell with a rubber septum. Raman scattering was excited at 406.7 nm with a Kr ion laser (BeamLok 2060, Spectra-Physics, Mountain View, CA) and 363.8 nm with an Ar ion laser (Spectra Physics, BeamLok 2080). The excitation light was focused into the cell at a laser power of 5 mW for the Fe(III) complex. Raman spectra were detected with a liquid N2-cooled CCD detector (Rober Scientific, Spec-10:400B/LN, Trenton, NJ) attached to a SPEX750M single polychromator (Jobin Yvon, Longjumeau, France). Raman shifts were calibrated with indene, acetone, CCl4, and an aqueous solution of ferrocyanide. 2.6. ESR spectra The Fe(III) hemin-bound KI (200 lM) complex was prepared in 50 mM Tris–HCl, pH 8.0, containing 50% glycerol at 25 °C. ESR spectra were recorded on a JEOL FE3X spectrometer (Tokyo, Japan) at 30 K. The magnetic field was calibrated using an NMR gauss meter (Echo Electronics, Hadsund, Denmark; model EFM-2000), and the temperature controlled with an Oxford 900 cryosystem [7,8]. 3. Results and discussion 3.1. Protein expression and purification At the beginning of this study, we attempted to express N-terminal His-tagged KI. Unfortunately, we could not obtain sufficient amounts of the soluble protein, in contrast to the report by Rafie-Kolpin et al. [9]. Therefore, the GST tag was employed to enhance protein solubility and avoid any interactions between His tag and heme. PreScission protease was applied to remove the GST tag. Fe(III) heme reconstitution was performed to ensure a 1:1 stoichiometry of KI protein:heme after purification. The expression efficiency and solubility of the Cys385Ala mutant were very low. Consequently, we only examined the spectroscopic characteristics of wild-type and Cys385Ser mutant KI proteins.
2.4. Optical absorption spectra
3.2. Optical absorption spectra of the KI domain
Optical absorption spectra were collected using Shimadzu UV-2500 and Shimadzu Multi Spec 1500 spectrophotometers (Kyoto, Japan) maintained at 25 °C. Spectral changes were recorded under both aerobic and anaerobic conditions in 50 mM Tris–HCl, pH 8.0, at 25 °C. The reaction mixture was incubated for 10 min prior to spectroscopic measurements to ensure stable solution temperature and proper coordination between Fe(III)–hemin or Fe(II)– heme and the protein. Experiments were performed at least three times for each complex.
We examined the optical absorption spectra of the heme iron-bound KI domain. The Soret band of wild-type KI with Fe(III) hemin was located at around 373 nm with a small shoulder at 412 nm (Fig. 2a; Table 1). The bandwidth was broader than that of full-length wild-type HRI enzyme. In view of these results, we suggest that the spectrum contains at least two species. The same peak around 370 nm was observed for the heme complexes of iron regulatory protein 2, another heme sensor protein [15], suggesting that the band around 370 nm is not due to a
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
Fe(III) Fe(II) Fe(II)CO Fe(II)NO
Absorbance
0.3 389 373
0.2
420 422
527
539 559
536 539 569
0.1 0
b 0.3 Absorbance
a
Fe(III) Fe(II) Fe(II)CO Fe(II)NO
389 420 391
0.2
1175
535 558 614
420
0.1
538
564
0.0
300
400
500
600
700
300
Wavelength (nm)
400
500
600
700
Wavelength (nm)
Fig. 2. Optical absorption spectra of heme-bound isolated wild-type (a) and Cys385Ser mutant (b) KI complexed to Fe(III) (bold solid line), Fe(II) (thin solid line in gray), Fe(II)–CO (dotted line) and Fe(II)–NO (broken line in red). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 1 Optical absorption spectra (nm) of the heme-bound full-length and isolated kinase insertion (KI) domain HRI complexes Fe(III) hemin
Wild-type full-length Wild-type KI Cys385Ser KI a
Fe(II) heme
Fe(II) heme–CO
Fe(II) heme–NO
Soret
Visible
Soret
Visible
Soret
Visible
Soret
Visible
366, 421 373, 412a 391
544 539 614
426 422 420
532, 559 527, 559 558
421 420 420
538, 564 536, 569 538, 564
398 389 389
538 539 535
Shoulder.
non-specific binding of the Fe(III)–hemin to the KI domain. Also, note that the spectrum of the free Fe(III)– hemin complex has broad peaks at approximately 387 and 357 nm, corroborating that the 373 nm species is not the free Fe(III)–hemin complex. Upon reduction of the heme iron with sodium dithionite, a narrow Soret band ascribed to the Fe(II) heme-bound wild-type KI domain was observed at 422 nm (Fig. 2a; Table 1). The absorption spectrum of the Fe(II) heme–CO complex of wild-type KI contained a Soret peak at 420 nm, similar to full-length wild-type HRI enzyme. The absence of the 450 nm peak upon CO binding to the Fe(II) heme–KI proteins implies that the Cys residue is not coordinated to the Fe(II) heme iron, being different from cytochrome P450 proteins. Our spectral findings differ from those of Rafie-Kolpin et al. [9] who report Soret absorption at 415 nm for the Fe(III) KI domain. The differences between the two sets of spectral results may be attributed to the His-tagged protein used in their experiments, in place of our GST-tag free protein. Since Cys385 is the only cysteine residue in the KI domain, we generated a Cys385Ser KI mutant, and examined the corresponding optical absorption spectra. The absorption spectra of Fe(III) hemin- and Fe(II) hemebound Cys385Ser KI were distinct from those of the wild-type KI domain in that the Soret bands of the mutant proteins were located at 391 and 420 nm, respectively (Fig. 2b; Table 1). The data suggest that ligand switching occurs in the mutant protein. In contrast, the spectrum of the Fe(II)–CO-bound Cys385Ser KI domain was similar to that of wild-type KI. Since there are four His residues in the KI domain, it is possible that ligand exchange occurs as a result of heme reduction and/or mutation of Cys385.
The full-length wild-type HRI protein is a 5-coordinated Fe(II) heme–NO complex displaying Soret absorption maximum at 398 nm [7]. In our experiments, the Fe(II) heme–NO complex of the isolated wild-type KI domain displayed a similar Soret band at 389 nm, supporting the presence of the 5-coordinated Fe(II)–NO complex. Interestingly, the Cys385Ser KI domain exists as a 5-coordinated Fe(II)–NO complex with a Soret absorption peak at 389 nm, similar to the cystathionine b-synthase complex [14]. Thus, the surrounding environment of the heme binding site of the mutant appears to retain a favorable conformation for the 5-coordinated Fe(II)–NO complex, despite subtle structural changes as a result of the mutation. 3.3. Heme binding properties of the KI domain We examined the heme binding characteristics of the Fe(III) hemin and Fe(II) heme complexes of wild-type and Cys385Ser mutant KI proteins. Differences in the absorption spectral changes of wild-type KI domain caused by the addition of Fe(III) hemin disclose a clear change in linearity, and support a heme to protein stoichiometry of 1:1 (Fig. 3a and c). The difference spectral titration of Fe(III) hemin with the wild-type KI domain also displayed a distinct positive peak at 360 nm (Fig. 3a). A Soret peak at around 360 nm characteristic of the thiol-coordinated high-spin complex of Fe(III) hemin-bound protein was observed for Fe(III) hemin-bound N-terminal truncated HRI protein [8,15]. Our findings suggest that the Cys residue in wild-type KI is an axial ligand for the Fe(III) hemin-bound form. Absolute optical absorption spectral changes of Fe(II) heme caused by addition of the wild-type KI domain
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
ΔAbsorbance
a 0.5
c
0.4 0.3 0.2 0.1 0.0 300
400
500
600
700
ΔAbsorbance at 360 nm
1176
0.5 0.4 0.3 0.2 0.1 0.0 0.0
Wavelength (nm)
d
Absorbance
1.0 0.8 0.6 0.4 0.2 0.0 400
500
600
700
Absorbance at 426 nm
b
300
0.5
1.0
1.5
2.0
Fe(III) hemin / apo KI
0.8 0.6 0.4 0.2 0.0 0.0
0.5
1.0
1.5
2.0
Fe(ll) heme/ apo KI
Wavelength (nm)
Fig. 3. Differences in absorption spectral changes of wild-type KI (10 lM) caused by addition of Fe(III) hemin (1 lM each) (a) and absolute optical absorption spectral changes of wild-type KI (10 lM) caused by addition of Fe(II) heme (1 lM each) (b). Spectral titration plots (filled circles) monitored at 360 nm and 426 nm for Fe(III) hemin (c) and Fe(II) heme (d), respectively, reveal a 1:1 stoichiometry of heme:protein for both complexes. The open circles in (d) represent the increase in absorbance of the free Fe(II) heme.
the possibility that axial ligand exchange occurs from Cys to another residue is suggested. Titration experiments were conducted for the Cys385Ser KI mutant. We did not observe any distinct changes in linearity reflecting a 1:1 stoichiometry of Fe(III) hemin binding to the mutant protein (Fig. 4a and c). The peak at 360 nm characteristic of the thiol-coordinated high-spin complex of Fe(III) hemin-bound mutant protein was
Δ
Δ
additionally support a 1:1 stoichiometry of heme to protein, although the change in linearity is less clear, compared with titration of the Fe(III) hemin complex (Fig. 3b and d). We believe that Fe(II) heme binding is relatively weak, compared to that of Fe(III) hemin, particularly when the axial ligand is thiol. This is a reasonable assumption, since the thiol negative charge and relatively electron-sufficient Fe(II) (compared with Fe(III)) repel each other. Rather,
Fig. 4. Differences in absorption spectral changes of Cys385Ser KI (10 lM) caused by addition of Fe(III) hemin (1 lM each) (a) and absolute optical absorption spectral changes of Cys385Ser KI (10 lM) caused by addition of Fe(II) heme (1 lM each) (b). Spectral titration plots (filled circles) monitored at 360 nm and 421 nm for Fe(III) hemin (c) and Fe(II) heme (d), respectively, disclose that the stoichiometry of heme:protein is not 1:1 for both mutant complexes. Specifically, for Fe(III) hemin binding, no alterations in the linearity of absorption change are observed, whereas for Fe(II) heme binding, the stoichiometry of heme:protein is 1:2. The open circles in (d) represent the increase in absorbance of the free Fe(II) heme.
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
significantly lower than that of wild-type protein (Figs. 3a and 4a). The data collectively imply that Cys385 is an axial ligand of the Fe(III) hemin-bound wild-type KI domain. For Fe(II) heme binding to Cys385Ser KI, the heme to protein ratio was calculated as 1:2, based on spectral changes of the mutant protein. We assume that unusual or non-specific binding occurs in this case. Therefore, titration of the Fe(II) heme complex into the Cys385Ser mutant led to ambiguous results, confirming that Cys385 is the axial ligand for Fe(II) heme-bound wild-type KI (Fig. 4b and d). In total, 4 His and 2 Met residues exist in the KI domain that may bind to heme as axial ligands.
c). Accordingly, we propose that the Fe(III) hemin-bound KI domain contains a Fe–Cys bond, similar to the fulllength HRI protein. The resonance Raman spectra indicate that the Fe(III) hemin-bound KI domain is in equilibrium between the 5-coordinated Fe(III)–Cys and 6-coordinated water or OH–Fe(III)–Cys complexes. In general, the 6coordinate low-spin heme does not show Fe–Cys or Fe– His stretching modes. 3.5. Electron spin resonance (ESR) spectra of the Fe(III) hemin-bound KI domain
406.7 nm 363.8 nm
KI
Intensity (a.u.)
377
341
415
338 ν (Fe-Cys)
Intensity (a.u.)
1620 1627
1502
1553 1582
1490
1428 1442
λ ex = 406.7nm
WT
ν8
c
HRI, pH 8.0
1374
Intensity (a.u.)
WT
HRI, pH 8.0 λex = 363.8 nm
675
b
347
Resonance Raman spectra provide valuable information about the heme coordination structure [10–12]. The marker band, m3, at the high frequency region of the KI domain is located at 1490 cm1 with a small band at 1502 cm1 (Fig. 5a), implying that the 5-coordinate high-spin complex is the major species, with 6-coordinate low-spin complex as a minor species, consistent with absorption spectral findings. This finding is distinct from that for the full-length HRI protein where the band at 1502 cm1 ascribed to the 6-coordinated low-spin complex is the major species, with a minor band at 1490 cm1 attributed to the 5-coordinated high-spin complex (Fig. 5a). In the low frequency region, the band ascribed to m(Fe– Cys) was observed at 338 cm1 (with laser excitation at 363.8 nm) for the Fe(III) hemin-bound full-length wildtype HRI protein (upper spectra in Fig. 5b and c). Similar bands in this region were observed for the Fe(III) heminbound KI domain protein (lower spectra in Fig. 5b and
338 ν (Fe-Cys)
Crystal field analysis of the ESR spectra of Fe(III) hemin low-spin signals reveals tetragonality and rhombicity [7,8]. These ESR parameters are useful to identify the axial ligands of low-spin Fe(III) hemin complexes. Thus, axial ligands of the Fe(III) hemin-bound full-length wildtype HRI were identified as His/Cys, whereas those of the N-terminal truncated HRI protein were OH/Cys (Fig. 6c, right) [7,8]. Fe(III) hemin-bound wild-type KI contained a mixture of high-spin and low-spin complexes, and the low-spin complex gave g values of 2.43, 2.26 and 1.91 (Fig. 6a). ESR parameters of the low-spin complex revealed that OH/Cys are the axial ligands of the Fe(III) hemin-bound wild-type KI domain, similar to the N-terminal truncated HRI protein (Fig. 6c, right). ESR spectra of the Fe(III) hemin-bound Cys385Ser mutant of the KI domain contained a mixture of several species (Fig. 6b). Since the ESR parameters of the low-spin complex are located outside the region analyzed, it was not feasible to identify the axial ligands of the Fe(III) hemin-bound Cys385Ser mutant KI domain. Therefore, the coordination of Cys385 in the Fe(III) hemin-bound wild-type KI was further corroborated from the ESR spectra.
3.4. Resonance Raman spectra of the KI domain
a
1177
WT
KI
KI
1400
1500
1600
Raman Shift (cm-1)
1700
300
400
500
600
Raman Shift (cm-1)
700
320
340
360
Raman Shift (cm-1)
Fig. 5. High frequency (a) and low frequency (b) regions of resonance Raman spectra of Fe(III) heme-bound wild-type full-length HRI (upper spectra) and KI (lower spectra) proteins. Spectra of the low frequency region are expanded in (c). For both proteins, bands at 338–341 cm1 (when excited at 363.8 nm) are assigned to the Fe–S stretching band, while close to it m8 band is observed at 347 cm1 upon excitation at 406.7 nm. This observation suggests that one of the axial ligands is Cys for the KI domain protein, similar to that of full-length HRI. Note that only the 5-coordinated Fe(III)–Cys complex, but not the 6-coordinated low-spin heme, shows the Fe–Cys stretching mode. WT, full-length wild-type HRI; KI, isolated wild-type KI domain.
OH–
1.99 1.91
6.09
a
2.43 2.26
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
4.35
1178
Fe3+
Fe3+
S
S
Cys385
Cys385
373 nm High spin
~ 412 nm Low spin e-
100
200
300
400
Magnetic Field (mT)
O C
100
200
2.00 1.91
2.28
2.46
6.08
b
4.39
Fe2+
300
400
Magnetic Field (mT)
Rhombicity |R / μ|
c
0.60
CO
X' Fe2+
X
X
420 nm Low spin
422 nm Low spin
Fig. 7. Proposed coordination structures of the isolated KI domain of HRI. The thiol group of Cys385 would be bound the Fe(III) hemin complex, at least in the isolated KI domain. The Fe(III) hemin-bound isolated KI domain would be in equilibrium between the 5-coordinated high-spin Fe(III)–Cys (major form) and 6-coordinated low-spin water or OH–Fe(III)–Cys (minor form) complexes based on the resonance Raman and ESR spectra. The axial ligand, Cys385, of the Fe(III) complex would switch to another unidentified amino acid residue upon reduction in that the axial ligand at the proximal side for both the Fe(II) and Fe(II)–CO complexes is no more thiol of Cys.
His/His
0.55
KI WT
0.50
His/Cys
0.45
-
HO /Cys
0.40 0.35 3
4
5
6
Tetragonality |μ / λ| Fig. 6. ESR spectra (left) of the wild-type (a) and Cys385Ser (b) proteins of the isolated KI domain of HRI. Based on the crystal field diagram for Fe(III) low-spin complexes (c), it is suggested that the isolated wild-type KI protein has OH/Cys as axial ligands, in contrast to the full-length wild-type HRI enzyme where His/Cys are possibly the axial ligands. The areas surrounded by the solid and dashed rectangles and dotted circles contain His/Cys, His/His, and Cys/water or hydroxyl anion as the axial ligands, respectively. The axial ligands of Cys385Ser KI could not be identified, since the parameters are located outside the region examined (data not shown). ESR spectra at the high spin region of the both wildtype and Cys385Ser mutant proteins did not provide valuable information about their coordination structures. WT, full-length wild-type HRI; KI, the isolated wild-type KI domain.
4. Summary Based on the optical, resonance Raman and EPR spectral results of the wild-type and Cys385Ser mutant proteins of the isolated KI domain, we conclude that the heme iron binds to Cys385 as an axial ligand for the Fe(III) hemin
complex, at least in the isolated KI domain. Perhaps the Fe(III) hemin complex would be in the equilibrium between two species with the high- and low-spin states (Fig. 7). However, upon reduction of the heme iron, thiol of Cys385 would dissociated from the heme iron and unidentified amino acid residue(s) would then become the axial ligands for both the Fe(II) heme and Fe(II) heme– CO complexes (Fig. 7). Fig. 8 summarizes the proposed structures of the HRI proteins. In the full-length wild-type protein (WT), His in the N-terminal region and Cys from the Kinase Domain at the C-terminal region are the candidate axial ligands of the Fe(III) heme. In the N-terminal truncated mutant (D145), OH and Cys in the Kinase Domain appear to be the axial ligands of the Fe(III) heme for the 6-coordinated low-spin complexes. However, the major form would be probably the 5-coordinated high-spin complexes with only one axial ligand of Cys385 in the Fe(III) hemin complex of the isolated KI domain. On the other hand, we cannot exclude the possibility that isolation of the KI domain from full-length wild-type HRI protein causes global protein conformational changes. Conformational changes in the N-terminal domain were observed upon its isolation so that heme ligand switching occurs upon the isolation of the domain. Specifically, His/His were proposed as the axial ligands of the isolated N-terminal domain of HRI (Fig. 7, far right), in contrast to His/ Cys for the full-length HRI protein (Fig. 7, far left) [7,12,13]. Moreover, full-length HRI was converted from the hexameric to the trimeric state upon removal of the
K. Hirai et al. / Journal of Inorganic Biochemistry 101 (2007) 1172–1179
Cys
KD
Fe NTD NTD His
WT His/Cys
Cys KD Fe OH-
Δ145 —HO/Cys
KD Cys KI Fe Fe
1179 His Fe
NTD His
KI Cys
NTD His/His
Fig. 8. Proposed structures of HRI proteins. WT, full-length wild-type HRI protein; D145, N-terminal truncated mutant of WT [9]; KI, isolated KI domain; NTD, isolated N-terminal domain [12,13].
N-terminal domain [8]. Examination of the isolated wildtype KI domain revealed a higher oligomeric state than hexamer. The regions surrounding the KI domain may thus be important to maintain its distinct structure.
Dr. H. Kurokawa for help with the initial stages of the work and Dr. A. Tanaka for valuable discussions.
5. Abbreviations
[1] A.G. Hinnebusch, Nat. Struct. Mol. Biol. 12 (2005) 835–838. [2] R.C. Wek, H.Y. Jiang, T.G. Anthony, Biochem. Soc. Trans., Part 1 34 (2006) 7–11. [3] D. Ron, Science 313 (2006) 52–53. [4] J. Marx, Science 313 (2006) 1564–1566. [5] T.E. Dever, Cell 108 (2002) 545–556. [6] J.J. Chen, in: N. Soneberg, J.W.B. Hershey, M.B. Mathews (Eds.), Translational Control of Gene Expression, Cold Spring Harbor Laboratory, Cold spring Harbor, NY, 2000, pp. 529–546. [7] J. Igarashi, A. Sato, T. Kitagawa, T. Yoshimura, S. Yamauchi, I. Sagami, T. Shimizu, J. Biol. Chem. 279 (2004) 15752–15762. [8] M. Miksanova, J. Igarashi, M. Minami, I. Sagami, S. Yamauchi, H. Kurokawa, T. Shimizu, Biochemistry 45 (2006) 9894–9905. [9] M. Rafie-Kolpin, P.J. Chefalo, Z. Hussain, J. Hahn, S. Uma, R.L. Matts, J.J. Chen, J. Biol. Chem. 275 (2000) 5171–5178. [10] T. Uchida, E. Sato, A. Sato, I. Sagami, T. Shimizu, T. Kitagawa, J. Biol. Chem. 280 (2006) 21358–21368. [11] T. Tosha, N. Kagawa, T. Ohta, S. Yoshioka, M.R. Waterman, T. Kitagawa, Biochemistry 45 (2006) 5631–5640. [12] J. Igarashi, A. Sato, T. Kitagawa, I. Sagami, T. Shimizu, Biochim. Biophys. Acta 1650 (2003) 99–104. [13] H. Ishikawa, B.G. Yun, S. Takahashi, H. Hori, R.L. Matts, K. Ishimori, I. Morishima, J. Am. Chem. Soc. 124 (2002) 13696–13697. [14] S. Taoka, R. Banerjee, J. Inorg. Biochem. 87 (2001) 245–251. [15] H. Ishikawa, M. Kato, H. Hori, K. Ishimori, T. Kirisako, F. Tokunaga, K. Iwai, Mol. Cell 19 (2005) 171–181.
HRI
heme-regulated inhibitor or heme-regulated eIF2a kinase eIF2a the a subunit of eukaryotic initiation factor 2 WT full-length wild-type HRI KI domain kinase insertion domain of HRI (amino acids 241–406 for mouse) C385S the Cys385Ser mutant of the isolated KI domain Fe(III) hemin Fe(III)–protoporphyrin IX complex Fe(II) heme Fe(II)–protoporphyrin IX complex GST-tag glutathione-S-transferase-tag
Acknowledgements This work was supported, in part, by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to M.M. and T.S.) and a Postdoctoral Fellowship for Foreign Researchers to M.M. from the Japan Society for the Promotion of Science. We thank
References