Biochemistry 1991, 30, 8357-8365 Segel, I. H. (1976) in Biochemical Calculations, pp 309, John Wiley and Sons, New York. Ugurbil, K. (1985) J. Magn. Reson. 64, 207-219. Verhoeven, J., Kramer, P., Groen, A. K., & Tager, J. M. (1985) Biochem. J. 226, 183-192.
8357
Vignais, P. V. (1976) Biochim. Biophys. Acta 456, 1-38. Williamson, J. R., & Corkey, B. E. (1969) Methods Enzymol. 13,434-513. Zweier, J. L., & Jacobus, W. E. (1987) J. Biol. Chem. 262, 80 15-802 1.
Similarities in Structure between Holocytochrome b5 and Apocytochrome b5: NMR Studies of the Histidine Residues? Cathy D. Moore, Ousaima N. AI-Misky, and Juliette T. J. Lecomte* Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania 16802 Received April 1, 1991; Revised Manuscript Received June 13, 1991 ABSTRACT: The properties of the six histidine residues of apocytochrome bShave been investigated by using one- and two-dimensional proton NMR spectroscopy in order to probe the structure remaining after heme removal. Spectral assignments were arrived a t by analyzing proton NOE connectivities, comparing them to those observed in the holoprotein, and inspecting the X-ray structure of the latter species. Each histidine residue was studied for its pKa value, interaction with the relaxation agent copper nitrilotriacetic acid, and reactivity toward bromoacetic acid. The four histidines which are not coordinated to the iron atom in the holoprotein (His-1 5, -26, -27, and -80) display in the major conformer of the apoprotein the same characteristic properties as in the holoprotein. Three of them are involved in specific interactions with the rest of the structure: His-15 and His-80 participate in hydrogen bonds, and His-27 is influenced by the nearby C-terminal segment. His-26 is the most exposed to the solvent. His-63 and His-39, which are located in the heme binding site, have distinct pKa values; they are affected differently by the copper agent and exhibit comparable reactivity toward bromoacetic acid, albeit milder than that of His-26. The results show that the heme binding residues are clearly distinguishable by their physicochemical properties and that several elements of native holoprotein structure are in place in the apoprotein. It is proposed that the structural influence of the heme is localized and that the amino- and carboxy-terminal segments form a structural unit providing stability to the apoprotein and supporting a fluctuating, partially folded binding site.
T e heme-containing, water-soluble fragment of cytochrome b5 (cyt 4)’ is a small globular protein made of both a-helices and &sheet (Mathews et al., 1971). Its heme prosthetic group is buried in the protein matrix and is known to play an essential role in the definition and stabilization of the structure of the holoprotein (Huntley & Strittmatter, 1972). In order to probe this role, we are currently characterizing physicochemical properties of the apoprotein by nuclear magnetic resonance spectroscopy. We recently reported results demonstrating the presence in the apoprotein of a stable cluster of side chains which overlaps with the second hydrophobic core of the holoprotein (Moore & Lecomte, 1990). This structured region contains the only tryptophan residue, Trp-22, and two of the protein’s six histidine residues, His-15 and His-80. In the holoprotein, His- 15 and His-80 are found on opposite sides of the /3-sheet, form hydrogen bonds with backbone atoms, and are thought to stabilize helices I and VI, respectively (Mathews et al., 1979). The other four histidine residues are located at positions 26, 27, 39, and 63. His-26 and -27 are situated in a turn connecting strands 3 and 4; His-39 and His-63 are the two axial ligands of the iron atom and terminate two of the four helices forming the heme cavity. Thus, the six histidine residues are scattered throughout the protein and provide good markers of structural features. In particular, the two heme Supported in part by BRSG Grant SO7 RR07082-22 awarded by the Biomedical Research Support Grant Program of the National Institutes of Health and in part by Grant DK 43101 from the National Institutes of Health. * To whom correspondence should be addressed.
0006-2960/91/0430-8357$02.50/0
pocket histidine residues are expected to report on the state of the vacant binding site. The assignment of the six histidines in rat liver apocyt b, is readily achieved by applying two-dimensional proton NMR techniques. To probe the environment of each residue, acidbase and paramagnetic relaxation agent titrations were performed. Histidine modification studies were also undertaken to determine relative reactivity. The results can be directly compared to those obtained in the holoprotein: the pK,’s of the histidine residues for two different isotypes of cyt bS,beef and rabbit, are available (Altman et al., 1989) while beef liver cyt b5 has been titrated with the relaxation agent Cu(NTA)for the purpose of electron-transferstudies (Reid et al., 1987). Chemical modification of the histidines using DEP has also been examined (Konopka & Waskell, 1988a,b; Altman et al., 1989). In this paper, we discuss the properties of the histidine residues of apocyt b5 as they contribute to a detailed description of the apoprotein. The data continue to show a native holoAbbreviations: apocyt b,, apo form of the water-soluble fragment of cytochrome b,; CD, circular dichroic; DEP, diethyl pyrocarbonate; COSY, two-dimensional correlated spectroscopy; cyt b,, water-soluble fragment of cytochrome b,; DQF-COSY, double-quantum-filtered COSY; holocyt b5, holo form of the water-soluble fragment of cytochrome b,; NMR, nuclear magnetic resonance; NOE, nuclear Overhauser effect; NOESY, two-dimensional nuclear Overhauser spectroscopy; NTA, nitrilotriacetic acid; SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis;TOCSY, total correlation spectroscopy; 2 4 , two-quantum spectroscopy.
0 1991 American Chemical Society
8358 Biochemistry, Vol. 30, No. 34, 1991 protein-like environment for His-I 5 and for His-80, extend the similarity to the surroundings of His-26 and His-27, and indicate that the unliganded His-39 and His-63 have properties modulated by their environment. They point to a specific structural and thermodynamic role for the @-sheetand the two helices which are not part of the heme binding site. MATERIALS AND METHODS Materials. Materials for cell growth were obtained from Difco Laboratories. Cell lysis was completed with lysozyme, RNase A, DNase I, dithiothreitol, and p-toluenesulfonyl chloride, all purchased from Sigma. The columns used for purification were a DEAE-Sephacel ion exchanger (a Pharmacia product purchased through Sigma) and a Bio-gel P-30 sizing gel (Bio-Rad). Acrylamide, Coomassie brilliant blue dye, and sodium dodecyl sulfate were obtained from Hoefer Scientific Instruments; other materials needed for the SDSPAGE or the nondenaturing gels came from Aldrich or Sigma. Reagent-grade cupric sulfate pentahydrate and NTA were purchased from Fisher and Kodak, respectively. Bromoacetic acid, approximately 99% pure, was acquired from Sigma, as was the horse skeletal myoglobin used for the neutral heme extraction. Protein Preparation. The protein used is the soluble fragment of rat liver cyt b5 expressed from a synthetic gene in Escherichia coli, strain TB-I (Beck von Bodman et al., 1986; the cell line was generously provided by Dr. S . G. Sligar). The cells were grown in LB broth at pH 7.4 and 37 OC in 2-L shaker flasks for approximately 18 h. Cells were harvested by centrifugation, and the reported method for cell lysis and protein purification was followed (Beck von Bodman et al., 1986). The resulting protein appeared as a single band by of 5.65. The apoprotein SDS-PAGE and gave an A412/A280 was prepared by using the methyl ethyl ketone method of Teale (1959). Residual holoprotein accounted for less than 1% of the total protein as evaluated by UV-vis spectroscopy. Protein concentrations were evaluated by using €413 = 114 mM-l cm-' for the holoprotein and e280 = 10.6 mM-' cm-' for the apoprotein (Strittmatter, 1960). NMR Samples and NMR Methods. Sample concentrations were 1-2 mM for one-dimensional and 2-4 mM for two-dimensional experiments; exchangeable protons were removed by lyophilizing the protein from Z H 2 0once after allowing the protein to exchange overnight at room temperature. 'H NMR spectra were recorded on a Bruker AM-500 spectrometer operating in the quadrature mode at a proton frequency of 500 MHz. The probe temperature was maintained at 298 K. Double-quantum-filtered COSY spectra (Rance et al., 1983) and two-quantum spectra (Braunschweiler et al., 1983; Rance & Wright, 1986) were acquired according to the standard procedure; phase-sensitive NOESY spectra (Kumar et al., 1980; Bodenhausen et al., 1984) were recorded with a Hahn echo (Davis, 1989) and sine modulation in the tl domain (Otting et al., 1986). The mixing times ranged from 50 to 200 ms; spin diffusion is noted at mixing times of 150 ms and longer in the 4 mM samples. TOCSY spectra (Braunshweiler & Ernst, 1983; Rance, 1987) were recorded with mixing times of 40 and 75 ms using the DIPSI-2 pulse train (Shaka et al., 1988). Quadrature detection in the wl domain was implemented through the time proportional phase increment method (Drobny et al., 1979; Marion & Wiithrich, 1983). The carrier and decoupler frequencies were phase-coherent and placed on the 'H2H0 line in all two-dimensional experiments (Zuidenveg et al., 1986). The 90° transmitter pulse was 7.4-7.6 ps. Water elimination was achieved by presaturation of the resonance with a 0.75- or 1.2-s decoupler pulse.
Moore et al. DQF-COSY spectra were recorded with a spectral width of 7042 Hz in both dimensions. For the 2 4 spectra, a spectral width of 7042 Hz was used in the w2 dimension and 14084 Hz in the wldimension. The raw data matrices contained 2048 complex points ( t 2 ) X 512 (tl) points. For the NOESY and TOCSY spectra, a spectral width of 12 500 Hz was used in the w2 dimension and 7042 Hz in the w1 dimension, yielding a typical data matrix of 4096 complex points ( t 2 ) X 512 ( t l ) points. A total of 96 transients were taken for each tl value. Data processing was performed with the FTNMR program of Hare Research Inc. on a DEC Microvax 11. In the case of the 2 4 spectra, the last 400 points of the free induction decays were zeroed to improve the sensitivity, and the w2 base line was corrected with a third-order polynomial function. Phase-shifted sine-bell functions were applied in both dimensions. For the NOESY spectra, high-sensitivity transforms were achieved by applying a sine-bell window function with a shift of 22.5' or 45O and zeroing the final 3072 points prior to transformation of the t2 domain. After transformation, the base line of each file was brought to zero by adding an appropriate constant. The width of the o2domain was reduced to 6250 Hz by keeping the central 2048 complex points. The t , domain was zero-filled once to increase the resolution and filtered with a phase-shifted sine-bell window function. All spectra were processed and plotted in the phase-sensitive mode. Chemical shifts are referenced to the lHZHOline at 4.76 ppm. p H Titrations. Solutions of approximately 1 mM apocyt b5 and oxidized holocyt b5 were prepared as described above. The pH* was measured with an Ingold slim-body combination electrode and adjusted in an Eppendorf microcentrifuge tube after removal of the sample from the NMR tube; the pH* values reported here are uncorrected meter readings (Meadows, 1972). The pH* titrations were performed by using two identical samples beginning at approximately pH* 7; the samples were titrated with 2-2.5-pL aliquots of either 0.1 M 2HC1until the protein precipitated in the acidic region at approximately pH* 5 , or 0.1 M Na@H until pH* 9.88 was reached. Spectra were recorded every 0.2 pH unit. The pH* was measured before and after each NMR spectrum was acquired, and the values agreed to within f0.03 unit, except above pH* 9.1 where they deviated within h0.05 unit. The resulting chemical shift versus pH curves for the C*H and C'H were fitted by a nonlinear least-squares routine using the equation: 6i = 8His (8His+ - 8His)( 10n(p"pH)/ [ 1 10n(p"pH)])
+
+
where bHis represents the chemical shift of the neutral form (high-pH limit), aHis+is the chemical shift of the protonated form (low-pH limit), and n is the Hill coefficient (Markley, 1975). In only one case (His-39) was the fit improved by letting n be adjusted by the program; all other n values were therefore fixed to 1. Paramagnetic Relaxation Agent Titration. A 2.8 mM sample of apocyt b5 was prepared, and the exchangeable protons were replaced by deuterons as described above. The paramagnetic relaxation agent Cu(NTA)- was prepared by dissolving approximately 2.5 mg of CuS04.5H20 in 1 mL of 2H20;1.9 mg of NTA was added to the solution, which resulted in a 9.85 mM solution as determined by visible spectroscopy (egm = 62 M-' cm-'; Kirson & Bornstein, 1960). The pH* of the protein solution was adjusted to 6.23, and that of the copper complex was 6.1, In this titration, small aliquots of Cu(NTA)- were added to the protein solution, and the pH* was checked and adjusted if necessary before the NMR spectrum was acquired. The solution was titrated with the copper complex to a final concentration of about 1 mM, which
Histidines of Apocytochrome b5
Biochemistry, Vol. 30, No. 34, 1991 8359
Table I: Chemical Shift and pK, Values for the Histidine Residues of Oxidized Holocytochrome b, and Apocytochrome b5 chemical shifta PK, proton apo holo apob holob holoC(beef) holo' (rabbit) 8.62 h 0.0Id 15 € 1 8.07 7.89 8.54 h 0.02d 6.76 8.47 f 0.0Id 8.73 f 0.02d 15 6 2 6.62 8.50 f 0.07 8.68 f 0.09 8.47 8.38 IS ave 8.35 6.93 f 0.0Id 6.89 f 0.02d 26 € 1 8.1 1 26 62 7.00 7.04 6.87 f O.0ZdJ 6.85 f O.OZd 26 av 6.90 f 0.07 6.87 f 0.07 6.92 6.91 27 €1 7.78 7.77 5 < pK, < 6s 5 < pK, < 6g 6.86 S < pK, < @ 5 < pK, < 6g 21 6, 6.88 5 < pK, < 6 5 < pK, < 6 NA~ 6.44 27 av 39 € 1 8.48 NA 7.67 f 0.01' NA NA 7.69 f 0.02' NA 39 6, 6.84 7.68 f 0.07 NA NA NA 39 av NA 63 € 1 8.30 NA 7.36 h O.Old 63 6, 7.17 NA 7.35 f 0.02d NA 63 av 7.36 f 0.07 NA NA NA 80 € 1 7.58 7.54
