146
Biochimica et Biophysica Acta, 1038 (1990) 146-151 Elsevier
BBAPRO 33607
13C-NMR of Clostridium pasteurianum ferredoxin after reductive methylation of the amines using [13C]formaldehyde * Martin Gluck and William V. Sweeney Department of Chemistry, City University of New York, Hunter College, New York, N Y (U.S.A.) (Received 16 October 1989)
Key words: Ferredoxin; Methylation; Protein stability; Amine group
Clostridium pasteurianum 2(4Fe-4S) ferredoxin has been reductively methylated using [t3C]formaldehyde and sodium cyanoborohlY3d~ride. Lys 3 and the N-terminal alanine, the only amines in the protein, are both dimethylated by this procedure. C-NMR titration of the apo, oxidized and reduced modified ferrodoxin indicate that the lysine pK is slightly over 10 in all three forms of the protein. In contrast, the N-terminal alanine shifts from a pK of 7.7 in the apoprotein to greater than 9 in both the oxidized and reduced modified ferredoxin. The unexpectedly high pK observed for the N-terminus is consistent with the presence of an ion pair in both the oxidized and reduced native forms of the protein. The methylated ferrodoxin is considerably less stable than the native protein, indicating an important role for the amines in protein stability.
Introduction Clostridium pasteurianum ferredoxin is a low molecular weight protein which contains two 4Fe-4S centers. It has been extensively studied through a large number of spectroscopic and chemical techniques. It has 70% sequence homology to another clostridial-type ferredoxin from Peptococcus aerogenes [1,2] for which an X-ray structure has been published [3,4]. As is common for clostridial-type ferredoxins, it has a limited amino-acid composition which is rich in acidic residues. This ferredoxin has only two amines, Lys 3 and the N-terminal alanine. In this paper, the possibility that the amine pK values of C. pasteurianum ferredoxin are oxidation-state dependent, is examined. It is likely that the conformation of this protein changes with reduction, since tritium-exchange studies of Clostridium acidi-urici ferredoxin demonstrate a different conformation in that protein's oxidized and reduced states [5]. Further, the X-ray structure of P. aerogenes ferredoxin indicates that
* This work was supported by NSF Grant DMB-8126808 and by a grant from the City University of New York Professional Staff Congress Board of Higher Education Award Program. Correspondence: W.V. Sweeney, Department of Chemistry, City University of New York, Hunter College, 695 Park Avenue, New York, NY 10021, U.S.A.
the N-terminal amine appears to be hydrogen bonded to Asp 37 [3]. Because of its intrinsic pK value of approx. 8, the N-terminal amine could be particularly important to protein binding in the physiological pH region. 13C-labeled formaldehyde is used to reductively methylate the amine groups, and the pK a values of the modified amines are determined using 13C-NMR [6]. This technique has been used to modify a large number of proteins., including the fd gene 5 DNA-binding protein [7,8], lysozyme [9,10], ribonuclease A [11-13], a-lactalbumin [14] and concanavalin A [15-17]. Because the methyl groups are small, only small changes in physical and chemical properties are found after methylation, and only small alterations in the amine pK a values are observed [61.
Experimental procedures Urea was recrystallized before use in ferrodoxin reconstitution procedures. All other chemicals were reagent grade and obtained from standard suppliers, and were used without further purification. 13C- and 14Clabeled formaldehyde were obtained from ICN. Only 65% of the radioactivity in the [14C]formaldehyde was found to arise from formaldehyde, as determined by the method of Jentoft and Dearborn [6]. In contrast, the [13C]formaldehyde was essentially pure, as judged from its carbon N M R spectrum.
0167-4838/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
147
C. pasteurianum was grown and ferredoxin isolated according to the procedure of Rabinowitz [18]. A crude preparation of C. pasteurianum hydrogenase was obtained by suspension of anaerobically harvested cells in 1 m l / g 0.15 M Tris-HC1 buffer (pH 8.80) containing 3 mM sodium dithionite. To this suspension 1 mg of egg white lysozyme per ml was added, and the suspension was allowed to incubate for 90 rain. The lysed cells were then spun for 50 min at 9800 × g and incubated at 60 °C for 10 min. The mixture, containing pelleted cell fragments and the treated supernatant, was then centrifuged again at 9800 × g for 20 min, and the light-green supernatant was applied to a small Whatman DE-52 DEAE column previously equilibrated with 0.10 M potassium phosphate buffer containing 1 mM sodium dithionite (pH 7.5). The column was then washed extensively with 0.10 M phosphate/1 mM dithionite buffer containing 0.15 M NaC1 until the pass through was colorless. The hydrogenase was eluted with phosphate buffer containing 0.20 M NaC1. Apoferredoxin was obtained by the procedure of Hong and Rabinowitz [19]. Reductive methylation of apoprotein was accomplished using a modification of the procedure of Jentoft and Dearborn [6]. In 0.1 M Hepes buffer (pH 8.3), 0.3 mM apoprotein was made 12 mM in formaldehyde. After a 1 h incubation, a 3-fold excess of sodium cyanoborohydride was added as a solid. Ferrodoxin was reconstituted using a modification of published procedures [20]. Because of relatively poor yields obtained during protein reconstitution, a second method of reductive methylation was also used. Hydrogenase-reduced ferrodoxin was reductively methylated under anaerobic conditions in 0.10 M Hepes buffer (pH 8.5). Three sequential additions of formaldehyde and cyanoborohydride were made, with 24 h incubations between additions prior to dialysis. Typically 0.1 mM ferredoxin was used, with additions of an 80-fold excess of formaldehyde followed by a 4-fold excess of cyanoborohydride over formaldehyde. Following reductive methylation the protein was repurified by DEAE chromatography. Protein with a ratio of A39o nm/m280 nm between 0.75 and 0.79 was used. For several samples of modified native protein, modified apoprotein was present as a contaminant. Contaminating apoprotein was easily detected in the carbon JNMR spectra. The extent of modification was monitored using [14C]formaldehyde after correction for the presence of radioactive impurities [6]. Because the amount of ferrodoxin was determined from the absorbance at 390 nm, a correction was made for the small amount (usually 0-10%) of apoprotein present following DEAE chromatography. The fraction of apoprotein present in a sample of ferrodoxin can be estimated from the absorbance spectrum (unpublished data). Samples after treatment with formaldehyde and cyanoborohydride,
followed by DEAE chromatography, were typically found to contain greater than 3.6 mol of amine methyl groups per mol ferrodoxin. Reduction potential measurements were conducted as previously described [21]. For stability studies (22 ° C), aerobic samples of ferredoxin (0.156 M NaC1/0.1 M Tris-HC1 buffer) were placed in uncapped 10 mm pathlength cuvettes, and the decrease in absorbance at 390 nm was followed as a function of time using a Varian 219 UV-visible spectrometer. In all cases the decrease was approximately linear for at least 8 h, and the stability was estimated from the slope of absorbance per initial absorbance vs. time plots. Assays of ferrodoxin-linked reduction of cytochrome C were carried out at 27 °C according to Lode et al. [221. NMR spectra were obtained using a JEOL-GX400 FT spectrometer. [aac]Methanol was used as an internal reference standard for carbon spectra (49.405 ppm relative to TMS). Protein samples were 1 mM or less in 50 mM potassium phosphate/100 mM NaC1 buffer. All 13C-NMR samples contained approx. 33% D20 to provide a lock signal, and pH values for titrations reflect uncorrected pH meter readings. For 13C-NMR titrations the pH was adjusted directly in the NMR tube using a Radiometer 26 pH meter connected to a 3 × 180 mm Ingold combination glass electrode. The sample was reduced using a hydrogen/hydrogenase system directly in the NMR tube. It was incubated for approx. 1 h under a hydrogen atmosphere before the NMR titration was begun. Visually observed bleaching of the sample, differences in chemical shift, and observation of the alanine resonance at pH values less than 8 provided assurance that the sample was reduced. The sample was flushed with hydrogen during pH adjustment. EPR spectra were recorded using a Varian X-band V-4500 spectrometer fitted with a Heli-Tran liquid helium transfer system (Air Products). Circular dichroism spectra were obtained using a Jobin Yvon Mark V dichrograph. Results and Discussion
Methylation Reaction of the protein with formaldehyde in the presence of sodium cyanoborohydride leads to methylation of the amines. Reaction with formaldehyde preferentially leads to dimethylation, since reaction with a second equivalent of formaldehyde proceeds more rapidly than the first [6]. Indeed, extensive efforts to obtain a monomethyl amine were unsuccessful. The 13C-NMR spectrum of modified apoferredoxin is shown in Fig. 1. In the decoupled spectrum of the methylated apoferredoxin, the two peaks shown correspond to di[13C]methylamino resonances. Proton cou-
148
APOF|RREDOXIN
PPM
Fig. 1. Proton noise de,coupled 13C-NMRspectrum (20 o C) of methylated C. pasteurianum apoferredoxin.Conditions: 30 mg/ml in 0.10 M Tris buffer, 100% D20 (pH meter reading 10.7); 1000 accumulations, pulse repeat time 2.04 s.
pied spectra show quartets with an average coupling constant of approx. 142 Hz, consistent with previous results [13]. The resonances near 42 and 44 ppm, respectively, are characteristic of unprotonated unperturbed dimethylated N-terminal and dimethylated lysine resonance [10,12,13]. The chemical shift and titration behavior of the resonance near 44 ppm in both native and apoferredoxin are typical of dimethylated lysine, and therefore this peak in both spectra is assigned to N,Ndimethyl Lys 3. The peak near 42 ppm is assigned to the N,N-dimethyl Ala ] resonance. These assignments are consistent with p K values obtained from titration of the resonances, as discussed below. The corresponding monomethyl resonances are expected between 31 and 34 ppm, and no evidence of any monomethyl derivative is observed (Fig. 1). Modification reactions were typically performed using either apoferredoxin or hydrogenase-reduced native ferrodoxin. For native ferredoxin, only a small extent of methylation was found using the conditions of Jentoft and Dearborn [6]. For this reason, larger excesses of formaldehyde were used. Reduced protein was used because it was found that less protein denatured during methylation in this oxidation state. Hong and Rabinowitz previously reported difficulty for reacting the N-terminus of oxidized native C. acidi-urici ferrodoxin in a range of modifying reactions [23]. The relative unreactivity of both Lys 3 and the N-terminus is surprising because the X-ray structure of the extensively homologous ferrodoxin from P. aerogenes indicates that both residues should be on the surface of the protein [3,4].
Even in the apoprotein the N-terminal amine does not react completely, as shown by the relative areas of the lysine and alanine peaks in the decoupled carbon N M R spectrum of the apoprotein (Fig. 1). This differential intensity is not a result of varying Overhauser enhancements, as nearly the same relative intensities are seen in the coupled spectrum. Because the p K of the N-terminal alanine is lower than that of lysine, it was expected [6,14] that it should preferentially react with formaldehyde. The reason for the diminished N-terminal reactivity in the apoprotein is not clearly understood, nor is the general lack of reactivity exhibited by the oxidized form of the protein. A similar lack of reactivity was been noted for the N-terminal amine of the fd gene 5 DNA-binding protein [7,8].
Comparisons of native and modified ferredoxins To assess the effects of methylation on the structure of the protein, a variety of spectroscopic techniques were used. The UV-visible spectrum of oxidized native and modified reconstituted ferrodoxin (0.1 M TrisHC1/0.25 M NaC1 buffer, p H 7.4) were essentially identical as was the CD spectrum of oxidized or reduced samples of native and modified material (0.05 M potassium phosphate/0.1 M NaC1 buffer, pH 7.4), and the EPR spectra of the reduced proteins (0.1 M TrisHC1/0.1 M NaC1 buffer, p H 8.3) were also virtually identical. However, slight differences were observed in the 1H-NMR spectra of native and reconstituted ferredoxin(0.1 M deuterated potassium phosphate/0.1 M NaC1 buffer, p H 7.19). The general features of the spectra are similar, but there are small (0.01 to 0.20 ppm) shifts in some of the downfield cysteinyl B proton resonances [24]. These small shifts are likely to indicate minor but real changes in conformation about the ironsulfur centers as a result of methylation. As judged by a ferrodoxin-dependent cytochrome C reduction assay, the activities of methylated and native ferredoxin were the same within the approximate 10% uncertainty of the assay. Similarly, there is no significant differences in the reduction potential between native and modified reconstituted ferredoxin in the pH range of 6.7-8.2. In contrast, as measured by the rate of decrease of the absorbance at 390 nm, the modified ferrodoxin was appreciably less stable than native ferredoxin (Fig. 2). In aerobic solution the modified reconstituted ferredoxin was approximately one third as stable as native ferredoxin in the p H range examined (pH 7.5 to 8.5, 0.1 M Tris-HC1 buffer, 22 ° C). The N-terminal amine is probably hydrogen bonded tO Asp 39, anchoring it to the body of the protein. This is indicated by the X-ray structure of the homologous P. aerogenes ferrodoxin and confirmed by the titration data presented later in this paper. It is possible that in the native protein the amine forms two hydrogen bonds
149 1.000 :.,.7
Apoferredoxin
0.980
C3 0 . 9 6 0 @3 rO .,~ 0 . 9 4 0 Cb O~ rO .
Biochimica et Biophysica Acta, 1038 (1990) 146-151 Elsevier
BBAPRO 33607
13C-NMR of Clostridium pasteurianum ferredoxin after reductive methylation of the amines using [13C]formaldehyde * Martin Gluck and William V. Sweeney Department of Chemistry, City University of New York, Hunter College, New York, N Y (U.S.A.) (Received 16 October 1989)
Key words: Ferredoxin; Methylation; Protein stability; Amine group
Clostridium pasteurianum 2(4Fe-4S) ferredoxin has been reductively methylated using [t3C]formaldehyde and sodium cyanoborohlY3d~ride. Lys 3 and the N-terminal alanine, the only amines in the protein, are both dimethylated by this procedure. C-NMR titration of the apo, oxidized and reduced modified ferrodoxin indicate that the lysine pK is slightly over 10 in all three forms of the protein. In contrast, the N-terminal alanine shifts from a pK of 7.7 in the apoprotein to greater than 9 in both the oxidized and reduced modified ferredoxin. The unexpectedly high pK observed for the N-terminus is consistent with the presence of an ion pair in both the oxidized and reduced native forms of the protein. The methylated ferrodoxin is considerably less stable than the native protein, indicating an important role for the amines in protein stability.
Introduction Clostridium pasteurianum ferredoxin is a low molecular weight protein which contains two 4Fe-4S centers. It has been extensively studied through a large number of spectroscopic and chemical techniques. It has 70% sequence homology to another clostridial-type ferredoxin from Peptococcus aerogenes [1,2] for which an X-ray structure has been published [3,4]. As is common for clostridial-type ferredoxins, it has a limited amino-acid composition which is rich in acidic residues. This ferredoxin has only two amines, Lys 3 and the N-terminal alanine. In this paper, the possibility that the amine pK values of C. pasteurianum ferredoxin are oxidation-state dependent, is examined. It is likely that the conformation of this protein changes with reduction, since tritium-exchange studies of Clostridium acidi-urici ferredoxin demonstrate a different conformation in that protein's oxidized and reduced states [5]. Further, the X-ray structure of P. aerogenes ferredoxin indicates that
* This work was supported by NSF Grant DMB-8126808 and by a grant from the City University of New York Professional Staff Congress Board of Higher Education Award Program. Correspondence: W.V. Sweeney, Department of Chemistry, City University of New York, Hunter College, 695 Park Avenue, New York, NY 10021, U.S.A.
the N-terminal amine appears to be hydrogen bonded to Asp 37 [3]. Because of its intrinsic pK value of approx. 8, the N-terminal amine could be particularly important to protein binding in the physiological pH region. 13C-labeled formaldehyde is used to reductively methylate the amine groups, and the pK a values of the modified amines are determined using 13C-NMR [6]. This technique has been used to modify a large number of proteins., including the fd gene 5 DNA-binding protein [7,8], lysozyme [9,10], ribonuclease A [11-13], a-lactalbumin [14] and concanavalin A [15-17]. Because the methyl groups are small, only small changes in physical and chemical properties are found after methylation, and only small alterations in the amine pK a values are observed [61.
Experimental procedures Urea was recrystallized before use in ferrodoxin reconstitution procedures. All other chemicals were reagent grade and obtained from standard suppliers, and were used without further purification. 13C- and 14Clabeled formaldehyde were obtained from ICN. Only 65% of the radioactivity in the [14C]formaldehyde was found to arise from formaldehyde, as determined by the method of Jentoft and Dearborn [6]. In contrast, the [13C]formaldehyde was essentially pure, as judged from its carbon N M R spectrum.
0167-4838/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
147
C. pasteurianum was grown and ferredoxin isolated according to the procedure of Rabinowitz [18]. A crude preparation of C. pasteurianum hydrogenase was obtained by suspension of anaerobically harvested cells in 1 m l / g 0.15 M Tris-HC1 buffer (pH 8.80) containing 3 mM sodium dithionite. To this suspension 1 mg of egg white lysozyme per ml was added, and the suspension was allowed to incubate for 90 rain. The lysed cells were then spun for 50 min at 9800 × g and incubated at 60 °C for 10 min. The mixture, containing pelleted cell fragments and the treated supernatant, was then centrifuged again at 9800 × g for 20 min, and the light-green supernatant was applied to a small Whatman DE-52 DEAE column previously equilibrated with 0.10 M potassium phosphate buffer containing 1 mM sodium dithionite (pH 7.5). The column was then washed extensively with 0.10 M phosphate/1 mM dithionite buffer containing 0.15 M NaC1 until the pass through was colorless. The hydrogenase was eluted with phosphate buffer containing 0.20 M NaC1. Apoferredoxin was obtained by the procedure of Hong and Rabinowitz [19]. Reductive methylation of apoprotein was accomplished using a modification of the procedure of Jentoft and Dearborn [6]. In 0.1 M Hepes buffer (pH 8.3), 0.3 mM apoprotein was made 12 mM in formaldehyde. After a 1 h incubation, a 3-fold excess of sodium cyanoborohydride was added as a solid. Ferrodoxin was reconstituted using a modification of published procedures [20]. Because of relatively poor yields obtained during protein reconstitution, a second method of reductive methylation was also used. Hydrogenase-reduced ferrodoxin was reductively methylated under anaerobic conditions in 0.10 M Hepes buffer (pH 8.5). Three sequential additions of formaldehyde and cyanoborohydride were made, with 24 h incubations between additions prior to dialysis. Typically 0.1 mM ferredoxin was used, with additions of an 80-fold excess of formaldehyde followed by a 4-fold excess of cyanoborohydride over formaldehyde. Following reductive methylation the protein was repurified by DEAE chromatography. Protein with a ratio of A39o nm/m280 nm between 0.75 and 0.79 was used. For several samples of modified native protein, modified apoprotein was present as a contaminant. Contaminating apoprotein was easily detected in the carbon JNMR spectra. The extent of modification was monitored using [14C]formaldehyde after correction for the presence of radioactive impurities [6]. Because the amount of ferrodoxin was determined from the absorbance at 390 nm, a correction was made for the small amount (usually 0-10%) of apoprotein present following DEAE chromatography. The fraction of apoprotein present in a sample of ferrodoxin can be estimated from the absorbance spectrum (unpublished data). Samples after treatment with formaldehyde and cyanoborohydride,
followed by DEAE chromatography, were typically found to contain greater than 3.6 mol of amine methyl groups per mol ferrodoxin. Reduction potential measurements were conducted as previously described [21]. For stability studies (22 ° C), aerobic samples of ferredoxin (0.156 M NaC1/0.1 M Tris-HC1 buffer) were placed in uncapped 10 mm pathlength cuvettes, and the decrease in absorbance at 390 nm was followed as a function of time using a Varian 219 UV-visible spectrometer. In all cases the decrease was approximately linear for at least 8 h, and the stability was estimated from the slope of absorbance per initial absorbance vs. time plots. Assays of ferrodoxin-linked reduction of cytochrome C were carried out at 27 °C according to Lode et al. [221. NMR spectra were obtained using a JEOL-GX400 FT spectrometer. [aac]Methanol was used as an internal reference standard for carbon spectra (49.405 ppm relative to TMS). Protein samples were 1 mM or less in 50 mM potassium phosphate/100 mM NaC1 buffer. All 13C-NMR samples contained approx. 33% D20 to provide a lock signal, and pH values for titrations reflect uncorrected pH meter readings. For 13C-NMR titrations the pH was adjusted directly in the NMR tube using a Radiometer 26 pH meter connected to a 3 × 180 mm Ingold combination glass electrode. The sample was reduced using a hydrogen/hydrogenase system directly in the NMR tube. It was incubated for approx. 1 h under a hydrogen atmosphere before the NMR titration was begun. Visually observed bleaching of the sample, differences in chemical shift, and observation of the alanine resonance at pH values less than 8 provided assurance that the sample was reduced. The sample was flushed with hydrogen during pH adjustment. EPR spectra were recorded using a Varian X-band V-4500 spectrometer fitted with a Heli-Tran liquid helium transfer system (Air Products). Circular dichroism spectra were obtained using a Jobin Yvon Mark V dichrograph. Results and Discussion
Methylation Reaction of the protein with formaldehyde in the presence of sodium cyanoborohydride leads to methylation of the amines. Reaction with formaldehyde preferentially leads to dimethylation, since reaction with a second equivalent of formaldehyde proceeds more rapidly than the first [6]. Indeed, extensive efforts to obtain a monomethyl amine were unsuccessful. The 13C-NMR spectrum of modified apoferredoxin is shown in Fig. 1. In the decoupled spectrum of the methylated apoferredoxin, the two peaks shown correspond to di[13C]methylamino resonances. Proton cou-
148
APOF|RREDOXIN
PPM
Fig. 1. Proton noise de,coupled 13C-NMRspectrum (20 o C) of methylated C. pasteurianum apoferredoxin.Conditions: 30 mg/ml in 0.10 M Tris buffer, 100% D20 (pH meter reading 10.7); 1000 accumulations, pulse repeat time 2.04 s.
pied spectra show quartets with an average coupling constant of approx. 142 Hz, consistent with previous results [13]. The resonances near 42 and 44 ppm, respectively, are characteristic of unprotonated unperturbed dimethylated N-terminal and dimethylated lysine resonance [10,12,13]. The chemical shift and titration behavior of the resonance near 44 ppm in both native and apoferredoxin are typical of dimethylated lysine, and therefore this peak in both spectra is assigned to N,Ndimethyl Lys 3. The peak near 42 ppm is assigned to the N,N-dimethyl Ala ] resonance. These assignments are consistent with p K values obtained from titration of the resonances, as discussed below. The corresponding monomethyl resonances are expected between 31 and 34 ppm, and no evidence of any monomethyl derivative is observed (Fig. 1). Modification reactions were typically performed using either apoferredoxin or hydrogenase-reduced native ferrodoxin. For native ferredoxin, only a small extent of methylation was found using the conditions of Jentoft and Dearborn [6]. For this reason, larger excesses of formaldehyde were used. Reduced protein was used because it was found that less protein denatured during methylation in this oxidation state. Hong and Rabinowitz previously reported difficulty for reacting the N-terminus of oxidized native C. acidi-urici ferrodoxin in a range of modifying reactions [23]. The relative unreactivity of both Lys 3 and the N-terminus is surprising because the X-ray structure of the extensively homologous ferrodoxin from P. aerogenes indicates that both residues should be on the surface of the protein [3,4].
Even in the apoprotein the N-terminal amine does not react completely, as shown by the relative areas of the lysine and alanine peaks in the decoupled carbon N M R spectrum of the apoprotein (Fig. 1). This differential intensity is not a result of varying Overhauser enhancements, as nearly the same relative intensities are seen in the coupled spectrum. Because the p K of the N-terminal alanine is lower than that of lysine, it was expected [6,14] that it should preferentially react with formaldehyde. The reason for the diminished N-terminal reactivity in the apoprotein is not clearly understood, nor is the general lack of reactivity exhibited by the oxidized form of the protein. A similar lack of reactivity was been noted for the N-terminal amine of the fd gene 5 DNA-binding protein [7,8].
Comparisons of native and modified ferredoxins To assess the effects of methylation on the structure of the protein, a variety of spectroscopic techniques were used. The UV-visible spectrum of oxidized native and modified reconstituted ferrodoxin (0.1 M TrisHC1/0.25 M NaC1 buffer, p H 7.4) were essentially identical as was the CD spectrum of oxidized or reduced samples of native and modified material (0.05 M potassium phosphate/0.1 M NaC1 buffer, pH 7.4), and the EPR spectra of the reduced proteins (0.1 M TrisHC1/0.1 M NaC1 buffer, p H 8.3) were also virtually identical. However, slight differences were observed in the 1H-NMR spectra of native and reconstituted ferredoxin(0.1 M deuterated potassium phosphate/0.1 M NaC1 buffer, p H 7.19). The general features of the spectra are similar, but there are small (0.01 to 0.20 ppm) shifts in some of the downfield cysteinyl B proton resonances [24]. These small shifts are likely to indicate minor but real changes in conformation about the ironsulfur centers as a result of methylation. As judged by a ferrodoxin-dependent cytochrome C reduction assay, the activities of methylated and native ferredoxin were the same within the approximate 10% uncertainty of the assay. Similarly, there is no significant differences in the reduction potential between native and modified reconstituted ferredoxin in the pH range of 6.7-8.2. In contrast, as measured by the rate of decrease of the absorbance at 390 nm, the modified ferrodoxin was appreciably less stable than native ferredoxin (Fig. 2). In aerobic solution the modified reconstituted ferredoxin was approximately one third as stable as native ferredoxin in the p H range examined (pH 7.5 to 8.5, 0.1 M Tris-HC1 buffer, 22 ° C). The N-terminal amine is probably hydrogen bonded tO Asp 39, anchoring it to the body of the protein. This is indicated by the X-ray structure of the homologous P. aerogenes ferrodoxin and confirmed by the titration data presented later in this paper. It is possible that in the native protein the amine forms two hydrogen bonds
149 1.000 :.,.7
Apoferredoxin
0.980
C3 0 . 9 6 0 @3 rO .,~ 0 . 9 4 0 Cb O~ rO .
