The Effects of Chemical Modifications on the Reconstitution, Activity, and Stability of Clostridial Ferredoxin*

SUMMARY Native ferredoxin isolated from Clostridium acidi-urici can-not be acetylated and is not digested by carboxypeptidase However, apoferredoxin, free of iron and sulfide, can be acet-plated and the COOH-terminal alanine and glutamine residues are quantitatively removed by carboxypeptidase Ferredoxin derivatives can be reconstituted from both modified proteins. The following derivatives

ties, iron and sulfide content, and biological activity.
In order to obtain reconstitution of ferredoxin from apoferredoxin, the cysteine sulfhydryl residues of apoferredoxin must be maintained in the reduced form (5) or protected as the sulfur-mercury derivative (6). Studies by immunological methods (7) and the tritium-hydrogen exchange technique (7) have indicated that apoferredoxin differs in conformation from ferredoxin.
The ability of apoferredoxin to form ferredoxin is apparently an intrinsic property of apoferredoxin, and no enzyme factors appear necessary under conditions (5, 6) that are reasonably physiological. Modification of apoferredoxin therefore provides an ideal opportunity to study the effect of structural alteration of a polypeptide chain on the reconstitution reaction.
Modification at the COOH-terminal end was achieved by the quantitative removal of alanine and glutamine with carboxypeptidase A. This was possible because the carboxyl-terminal sequence of Clostridium acidi-urici ferredoxin is -Pro-Val-Gln-Ala-(3) and the specificity of carboxypeptidase A is such that it does not hydrolyze the prolyl-amino acid bonds. Chemical modification of the NHt-terminal amino group was feasible because apoferredoxin from this organism contains no other free amino group (3) and it can be prepared in the oxidized form (5) so that possible chemical reactions also affect the cysteine sulfhydryl groups are eliminated.
These apoproteins were previously referred to as apoferredoxin I, apoferredoxin II, and apoferredoxin III, respectively Dr. D. Levy, Mr. S. M. Shiigi, and Dr. F. H. Carpenter. Other chemicals were obtained from sources previously noted. C. acidi-urici was grown as previously described (9) except that sodium carbonate and sulfuric acid were omitted from the media.
Previously described procedures were used for the preparation of ferredoxin (4), apoferredoxinn,z (6), and apoferredoxinox (5) from this organism, as well as for the assay of ferredoxin in an enzymic test (10).

Methods
A&y&ion-The procedure used was based on the method described by . "C-Acetic anhydride (uniformly labeled) (26 ~1) was added in small increments over a period of 1 hour to apoferredoxin 0X or native ferredoxin (18 mg) in 4 ml of 50% saturated sodium acetate at O", and the reaction was allowed to continue for another 1.5 hours.
With the apoferredoxin, the mixture was then dialyzed for 30 hours at 4" with three changes of 0.01 M Tris-chloride buffer, pH 8.5 (3 liters each).
In the case of the ferredoxin sample, the mixture was passed over a Sephadex G-25 column equilibrated with 0.1 M Tris-chloride buffer, pH 7.4, and the column was developed with the buffer.
After 1 hour, another 50 mg of the reagent were added and the pH was adjusted to 8.5. An hour later, 50 mg of the reagent were added, the pH was adjusted to 8.4, and the reaction was allowed to proceed for another hour.
The protein was reisolated by gel filtration on a Sephadex G-25 column equilibrated with 0.01 M sodium acetate and the column was developed with the same solution.
Succinylation-The procedure used was based on the method of Klotz (13). ilpoferredoxin,, (73 mg, 13.2 pmoles) was dissolved in 13 ml of water and succinic anhydride (60 mg, 0.6 mmole) was added with magnetic stirring in small increments to the solution over 1 hour at room temperature.
The pH was maintained between 7 and 8 with addition of 3 N NaOH. The solution was allowed to stand for an additional 40 min and was then dialyzed overnight against 12 liters of 0.01 M sodium acetate.
The determination of 0-succinylation of hydroxyamino acid residues was carried out with the alkaline hydroxylamine reaction of Hestrin (14), whereas the procedure of Lipmann and Tuttle (15) was used for the estimation of active acyl groups with glycine methyl ester hydrochloride as standard. Synthesis of AminoacylapoferredoxGa~Aminoacylapoferredoxins were synthesized from apoferredoxin by the method of Levy and Carpenter (16). Apoferredoxin,, (80 mg, 14.4 pmoles) was dissolved in 8 ml of 25:75 water-dimethylformamide medium? and the p-nitrophenyl ester of a BOC-amino acid (170 pmoles) and 15 ~1 of triethylamine were added. The reaction was allowed to proceed, with magnetic stirring, for 20 hours at room temperature.
The solution was then passed over a Sephadex G-25 column (1.6 x 40 cm) to remove p-nitrophenol formed, 2 The abbreviations used are: BOC, t-butyloxycarbonyl group; 1>TNB, 5,5'-dithiobis (2.nitrobenzoic acid). 3 We thank Dr. I>. Levy,Mr. 9. M. Shiigi,and Professor F. H. Carpenter for making available to us the method for aminoacylation in aqueous dimethylformamide. and the cloudy, excluded fractions were pooled and extracted three times with equal volumes of ether.
The aqueous portion was lyophilized under high vacuum. The lyophilized material was placed in a glass centrifuge tube and was dried thoroughly over PZO~ under high vacuum for 20 hours.
The solid material was then dissolved in 2 to 3 ml of anhydrous trifluoroacetic acid and allowed to stand for 1 hour at room temperature.
At the end of that time the solution was cooled in an ice bucket and the protein was precipitated with 4 ml of cooled ether. The precipitate was washed twice with ether and dried under nitrogen gas. The white solid material was dissolved in 5 ml of 0.1 M Tris-chloride buffer, pH 8.5, with a small amount of 1 M NaOH added to neutralize the amount of trifluoroacetic acid, and the solution was centrifuged and the supernatant was dialyzed against 0.01 M sodium acetate overnight.
The yield was about 75 to 90%.
For the synthesis of glutamylapoferredoxin, saponification was carried out before the treatment with trifluoroacetic acid. After lyophilization, the solid material was dissolved in 10 ml of 0.034 M sodium carbonate buffer, pH 11.0, and allowed to stand 3 hours at room temperature (16). The protein was then precipitated by the addition of a solution of 30% trichloracetic acid to bring the final concentration to 5% and the precipitate was washed with water and lyophilized.
Carboxypeptidase A Treatment-Apoferredoxin,, (50 mg) in 17 ml of 0.1 RI Tris-chloride buffer, pH 7.6, was treated with 2 mg of carboxypeptidase A at room temperature for 7 hours. The enzyme solution was prepared by diluting 1 volume of the commercial stock suspension with 10 volumes of 10% LiCl and stirring for 1 hour at 0". The concentration of enzyme was determined from absorbance at 278 nm (17). Ferredoxin was treated under the same conditions but was under anaerobic conditions to prevent the deterioration of ferredoxin. At the desired interval a 0.2-ml aliquot was withdrawn to 1 ml of 0.2 M sodium citrate buffer, pH 2.2, mixed, and centrifuged. The supernatant was then analyzed for amino acids on a Beckman-Spinco automatic amino acid analyzer, model 120 (18). lodination-Iodination was performed by the procedure of Gruen,Laskowski,and Scheraga (19). Apoferredoxin,, (41 mg, 7.4 pmoles) was dissolved in 5.5 ml of 0.5 M glycine-NaOH buffer, pH 9.5, and was iodinated over a period of 2.5 hours at 0" with 1 ml of 0.14 M '311-triiodide solution (1 Hmole = 60,900 cpm) (19). The protein was separated from the reagent by gel filtration on a Sephadex G-25 column previously equilibrated with 0.01 M Tris-chloride buffer, pH 8.5, and the column was developed with the same buffer.
Preparation of BOC-apojerrecloxin-BOC-apoferredoxin was synthesized from apoferredoxin by the method of Levy and Carpenter (16). Apoferredoxin,, (70 mg, 11.3 pmoles) was dissolved in 10 ml of 25:75 water-dimethylformamide medium and 180 mg of BOC-azide and 25 ~1 of triethylamine were added. The solution was incubated for 7 hours at 40" and then the protein was reisolated by gel filtration on a Sephadex G-25 column (1.6 x 45 cm), equilibrated with water, and lyophilized. The yield was 87%.
Regeneration of Apojerredoxin from BOC-apoferredoxin-The condition used was the same as that used for the removal of the blocking group from the BOC-aminoacylapoferredoxins. The regenerated apoferredoxin was obtained in 74% yield. Determination of Free Amino Groups--The degree of modification of the free amino groups was determined by measuring the decrease in the ninhydrin color given by the protein (11). The per cent of amino groups modified was calculated from the difference between the free amino groups present in the original and the modified proteins.
Reconstitution of Ferreobxin Derivatives-The procedure used for reconstitution of ferredoxin derivatives from the modified apoferredoxin,, derivatives is the same as that previously described for apoferredoxin,, (initial A 390 about 1) were kept in 0.1 JI Trischloride buffer, pH 7.4, containing 0.1 M NaCl and decrease in AZ90 was followed at 4". All samples were centrifuged in a Siorvall centrifuge at 4" for 10 min at 9900 X g before measuring absorption.
The decrease in A~90 was expressed as the per cent of the initial Aag0 but has been corrected for the absorbance due to the presence of an amount of unmodified ferredoxin (18% in acetyl-, 207, in acetimido-, and 177, in BOC-ferredoxin) in the rcconst,ituted ferredoxin derivatives. A, native ferredoxin, A; R, acetimidoferredoxin, E; C, acetylferredoxin, 0; and D, BI)C-ferredoxin, 0.
purified by precipitation with ammonium sulfate. This procedure, involving reconstitution, isolation, and purification of the ferredoxin derivatives, required roughly 1.5 hours.
The stability of the purified ferredoxin derivatives was determined by incubating the isolated and purified product in 0.1 RI Tris-chloride, pH 7.4, containing 0.1 M NaCl at 4" under aerobic conditions. At intervals, aliquots were withdrawn and assayed for activity in the phosphoroclastic reaction (10).

RESULTS
Acetylferredoxin-The extent of the acetylation of ferredoxin and the apoferredoxins with 14C-acetic anhydride was determined from the radioactivity of the reisolated protein.
When native ferredoxin was subjected to acetylation with acetic anhydride, only 0.07 Fmole of 14C-acetyl group was incorporated into 1 pmole of protein.
Although the NHZ-terminal group of native ferredoxin does not react with acetic anhydride, apoferredoxinox incorporated 0.85 pmole of 14C-acetyl group per pmole of protein (Table I).
The reconstituted acetylferredoxin contained 0.82 pmole of 14C-acetyl group per pmole of protein (Table I), and was obtained from apoferredoxin,, in a yield of 54%.
Acetylferredoxin had spectral properties identical with those of native ferredoxin.
However, it was less stable than native ferredoxin as determined by its activity in the phosphoroclastic reaction (Fig. 1, Curve B). The decrease in activity followed first order kinetics with a half-time of 11 hours under the condi- tions used. Under these conditions, native ferredoxin has a half-life of greater than 500 hours. In addition to the decrease of the biological activity, the absorption of acetylferredoxin at 390 nm was also found to decrease with time. The decay of the chromophore also followed first order kinetics (Fig. 2, Curve C). Acetimidoferre&xicetimidoferredoxin was obtained in crystalline form (Fig. 3). The ninhydrin reaction showed that approximately 20% of the amino groups of the acetimidoapo-ferredoxinoX and reconstituted acetimidoferredoxin were free (Table I). The absorption spectrum of the reconstituted acetimidoferredoxin was indistinguishable from that of native ferredoxin. The derivative was 78% as active as native protein in the enzyme assay. This derivative is more stable in solution  Fig. 1 (Curve A) and the rate of loss of its absorption at 390 nm is shown in Fig. 2 (Curve B). The half-time for the decay of activity was 132 hours. BOC-ferredoxin-The absorption spectrum of the reconstituted BOC-ferredoxin was indistinguishable from native ferredoxin. It was about 78% as active as the native ferredoxin in the enzyme test. Approximately 14 and 17%, respectively, of the amino group of BOC-apoferredoxin,, and BOC-ferredoxin was unblocked as determined by the ninhydrin reaction (Table I). It was not as stable as acetylferredoxin.
The half-time for the activity decay was about 4.5 hours. The decay of the activity and the decay of the absorption at 390 nm are shown in Fig. 1 (Curve C) and Fig. 2 (Curve D), respectively. The reconstituted BOC-ferredoxin was obtained from BOC-apoferredoxin,,d in about 43% yield. The BOC-ferredoxin was not obtained in crystalline form.
They were obtained in about 75 to 90% yield. Ferredoxin derivatives could be reconstituted from these aminoacylapoferredoxin,,d derivatives and were obtained in 45 to 54% yield. The aminoacylferredoxins were obtained in crystalline form. The photomicrographs of these crystals are shown in Fig. 3. Amino acid analysis on these reconstituted aminoacylferredoxin derivatives showed that a single amino acid residue had been added to the protein.
The results of the amino acid analysis are shown in Table II. The addition of methionyl, phenylalanyl, or lysyl residues to the protein was apparent from the amino acid analysis because the native protein does not contain any residues of these amino acids.
The reconstituted aminoacylferredoxins were less active and  (initial Aa90 about 1) were kept in 0.1 M Tris-chloride buffer, pH 7.4, containing 0.1 M NaCl and the decrease in AWI was followed at 4". Samples were centrifuged in a Sorvall centrifuge at 4" for 10 min at 9900 X g before measuring absorption. The decrease in A390 was expressed as per cent of the initial AW. A, native ferredoxin, A; B, glycylferredoxin, q i; C, lysylferredoxin, n ; D, phenylalanylferredoxin, l ; E, methionylferre- retain the same net charge as native ferredoxin. The stability data show that the derivatives in which an amino acid residue with relatively small R groups were introduced were more stable than those with larger ones. Glutamyl-and lysylferredoxins represent derivatives in which one negatively or one positively charged side chain, respectively, was added to the native protein.
Glutamylferredoxin was 2.5 times less stable than lysylferredoxin.
The stability of phenylalanylferredoxin crystals was also examined.
When stored as crystals suspended in ammonium sulfate (70% saturation) in 0.15 M Tris-chloride buffer, pH 7.4, at 4" under aerobic conditions, this derivative was found to retain 57.3% of the specific biological activity of the native protein after 41 days of storage, a loss of 16% activity.
This result indicates that ferredoxin is more stable when stored as a suspension of crystals than when stored in solution.
Regenerated Ferredoxin from BOC-apoferredoxin-Treatment of BOC-apoferredoxin with anhydrous trifluoroacetic acid resulted in the regeneration of an apoferredoxin which could be converted to ferredoxin and crystallized (Fig. 5). The reconstituted ferredoxin was indistinguishable from the native ferredoxin with respect to the spectral properties, biological activity, and the stability as determined by its biological activity and absorption at 390 nm. Des-(Ala55,Gln64)-Ferredoxin-Treatment of apoferredoxinn, with carboxypeptidase A resulted in the removal of alanine and glutamine from the carboxyl-terminal end of the protein.
The time course of the digestion is shown in Fig. 6. It is apparent that the digestion was essentially complete in 2 hours. NO other free amino acids were detected even after 7 hours of digestion. Ferredoxin derivative was reconstituted from the des-(Ala55, Gln54)-apoferredoxin,, by addition of iron and sulfide according to the conditions described by Malkin and Rabinowitz (6), and was obtained in 58% yield. The reconstituted ferredoxin derivative was indistinguishable from the native ferredoxin with respect to spectral properties.
It was about 78% as active as the native ferredoxin (Table III), but was not as stable as the native ferredoxin. Fig. 7 (Table III). The reconstituted des-(Ala55,Gln5~)-ferredoxin contained 1 alanine and 1 glutamic acid residue less than native ferredoxin.
Native ferredoxin is not digested by carboxypeptidase A under conditions in which apoferredoxin,, is digested. This result suggests that the COOH-terminal amino acid residues of the native ferredoxin are not accessible to the carboxypeptidase A and might be buried in the native structure.
SuccinylapofemedoxiSuccinylation of apoferredoxin,, with succinic anhydride resulted in the loss of 83% of the free amino group as determined by the ninhydrin reaction (Table I). The succinylapoferredoxin,, was converted to succinylapoferredoxin,,d.
Since succinic anhydride has been shown to react, not only with amino group but also with the side chain -OH groups of tyrosine, serine, and threonine residues in proteins (20), this inability of succinylapoferredoxin to form ferredoxin derivative might possibly be due to the succinylation of -OH groups of these amino acid residues in apoferredoxin.
Succinylapoferredoxin was analyzed for the presence of 0-succinyl groups.
The result of the alkaline hydroxylamine reaction of He&in (14) showed the presence of approximately one 0-succinyl per mole of protein whereas no active acyl group was found by the method of Lipmann and Tuttle (15). These results indicate that one -OH group of serine or threonine or a combination of these 2 amino acid residues in succinylapoferredoxin may have been succinylated. Tetratidoapojerredoxin-Approximately 3.6 atoms of 1311 were incorporated into 1 mole of apoferredoxin,,. This is slightly lower than 4, a value expected for iodination of the 2 tyrosine residues in the protein.
A ferredoxin derivative could not be reconstituted from the iodoapoferredoxin,,d. Sulfhydryl group analysis of the iodinated apoprotein by the DTNB method (21) after reduction with NaBH4 in 8 M urea anaerobically showed that, the protein had 7.1 sulfhydryl groups per mole of protein. This result indicates that the disulfides of apoferredoxin,, were not oxidized to the sulfonic acid level by iodination, although the formation of other oxo-derivatives that would be reduced to sulfhydryl level by borohydride was not eliminated, and that the The derivative was assaved for the activity in the phosphoroclastic reaction. The activity"of the derivative was expressed as per cent of that of the native ferredoxin.
The decrease in the absorption at 390 nm was measured as described in Fig. 2. inability of iodoapoferredoxin,,d to reconstitute may most probably be attributed to the modification of the tyrosine residues. When ferredoxin was iodinated under conditions identical with those used for iodination of apoferredoxin,,, it was bleached, indicating that iron and acid-labile sulfide were removed from the protein.
Ferredoxin is apparently too labile to be subjected to iodination.