The Primary Structure of the Clostridium tartarivorum Ferredoxin, a Heat-stable Ferredoxin*

SUMMARY The amino acid sequence of the Ctostridium tartarivorum ferredoxin was determined to be: Ala-His-Ile-Ile-Thr-Asp-Glu-Cys-Ile-Ser-Cys-Gly-Ala-Cys-Ala-Ala-Glu- Cys -Pro -Val-Glu-Ala-Ile-His-Glu-Gly-Thr-Gly-Lys-T~-Gln-Val- Asp-Ala-Asp-Thr-Cys-Ile-Asp-Cys-Gly-Ala-Cys-Gln-Ala-Val-Cys-Pro-Thr-Gly-Ala-Val-Lys-Ala-Glu. The tirst 23 residues were determined in the protein sequenator while the remainder of the sequences were determined by con-ventional methods. The primary structure of the heat-stable ferredoxin is compared with the ferredoxins from the mesophilic organisms.


Sequence of a Heat-stable
Ferredoxin Vol. 246,No. 12 Begg was utilized (14). The ethyl acetate-soluble PTH-amino acids were identified by gas-liquid chromatography in the Beckman model GC-45 gas chromatograph.
The two columns used and the experimental conditions for separating the PTH-amino acids or the trimethylsilyl derivative of the PTH-amino acids have been described by Pisano and Bronzert (15,16). The PTH-histidine which remained in the acidic aqueous layer after ethyl acetate extraction was extracted after adjusting the pH to 8 and was quantitated on the automatic amino acid analyzer after acid hydrolysis of the samples in 6 N HCl for 18 hours.
Hydrolysis of Ae-jerredoxin by Trypsin-The Ae-ferredoxin (16 mg) was incubated with TPCK-trypsin (enzyme to substrate was 1:20) at pH 8.0 in a total volume of 1.6 ml.
Additional trypsin was added at 12 and 24 hours and the reaction was performed at 28" for 36 hours. Chymotryptic Digestion of Ae-ferrecloxin-In this case the protein (27 mg) was reacted with TLCK-chymotrypsin (enzyme to substrate was 1:50) at pH 8.0 in a total volume of 2.7 ml. The reaction time was 5 hours and the temperature 40". Thermolysin Hydrolysis of Peptide T-4--Peptide T-4 (0.7 pmole) was reacted with thermolysin (enzyme to substrate was 1: 20) at pH 8.0 in a volume of 1.0 ml.
The hydrolysis reaction was carried out at 40" for 10 hours.
Pepsin Digestion of Peptide T-B-About 0.5 pmole of the peptide was reacted with pepsin (enzyme to substrate was 1: 20) in 1.0 ml of 0.01 N HCI.
The hydrolysis time was 18 hours and the temperature was 40". AG l-X2 Column Chromatography of Tryptic Digest. of Aejerredozin-The digest was applied to a column of AG l-X2 (0.7 X 50 cm).
(b) Tubes 119 to 221, the gradient elution was obtained by mixing 100 ml of the pH 6.5 buffer mentioned above with 100 ml of 0.5 M acetic acid.
The fractions were assayed for peptide by the ninhydrin assay of Moore and Stein (17).
AG l-X2 Column Chromatography of Thermolysin Digest of Peptide T-&--The digest was applied to a column (0.7 x 20 cm). For tubes 1 to 10, the buffer was 0.1 M pyridine-acetate buffer, pH 7.0. For tubes 11 to 25, the buffer was 0.1 M pyridineacetate buffer, pH 5.0. The columns were monitored by the ninhydrin assay method (17). Nomenclature-Peptides obtained hydrolysis with trypsin and chymotrypsin are denoted by "T" and "C," respectively. Products of further hydrolysis of these peptides with thermolysin and pepsin are denoted by "Th" and "P," respectively.

Determination of NHz-terminal Sequence
The NHz-terminal sequence of the Cm-ferredoxin from C. tartarivorum was determined in a Beckman protein sequencer. Four separate runs were made in which 0.2 to 0.7 pmole of protein was taken for analysis.
The amount of protein did not have much effect on the number of sequences which could be determined in the instrument. A typical result from one of the runs is shown in Fig. 1 Table I Table II. mental section for details.

Sequence of Peptides
As mentioned earlier, the first 23 residues from the NH2terminal end of the protein were determined by the protein sequencer.
This left 32 residues the sequence of which remained undetermined.
Although additional sequence studies were performed on fragments from the first 23-residue portion, they will be omitted here.
In order to complete the sequence, the peptide was hydrolyzed with thermolysin.
Two peptides (T-4-Th-1 and T-4-Th-2) were isolated by column (Fig. 4) chromatography of the digest and the amino acid composition and other properties of the peptides are summarized in Table III.
drolyzed by pepsin and two fragments were obtained by paper chromatography.
The amino acid composition and properties of T-5-P-l and T-5-P-2 are summarized in Table VI. Hydrazinolysis was performed on Peptide T-5-P-l (Tyr-Gln-Val-Asp) and aspartic acid was obtained in 80% yield which in combination with sequence data on T-5 itself, completed the sequence of this peptide.
Table VII summarizes the sequence data on Peptide T-5-P-2 which was shown to be Ala-Asp-Thr-Cys(Ae) (residues 34 to 37).   Peptide T-6: Ile-Asp-Cys(Ae)-Gly-Ala-Cys(Ae) (Residues 58 to @)-The sequence data of this peptide are summarized in Table VIII. Peptide T-7: Gln-Ala-Val-Cys(Ae)-Pro-Thr-Gly-Ala-Val-Lys (Residues 4.4 to 6S)-The sequence data obtained on Peptide T-7 are summarized in Table IX. Peptide  Sequence: Ile-Asp-Cys(Ae)-Gly-Ala-Cys(Ae) -t-r---  The end groups of the chymotryptic peptides were determined in order to identify the end groups and to check the purity of the samples. Since it is possible to place all the tryptic peptides by analogy to the sequence of the ferredoxins which have been determined, the details will be omitted here.

Reconstruction
of the amino acid sequence of Clostridium lartarivorum ferredoxin from the sequenator, tryptic peptide, and chymotryptic peptide data.
Peptides T-l, T-2, T-3, and T-4 have the amino acid composition which corresponds to the sequence. Peptides C-l, C-2, and C-3 also have the amino acid composition which corresponds to this sequence.
Peptide T-5 is adjacent to T-4 since Peptide C-3 overlaps these two peptides.
Peptide T-6 is adjacent to Peptide T-5 since Peptide C-4 overlaps these two peptides.
Peptide T-7 is adjacent to Peptide T-6 since Peptide C-7 contains Peptides T-7 and T-8 and since C-7 and the protein itself have been shown by hydrazinolysis to contain COOH-terminal glutamic acid residues. The yield of glutamic acid from Peptide C-7 was about 90%.
Of course, it would have been also possible to place the tryptic peptides by analogy to the sequences of other ferredoxins which have been already determined.
The complete sequence of the heat-stable ferredoxin from C. tartarivorum is shown in Fig. 5. The amino acid composition, heat stability, optical rotatory dispersion curves, molecular weight, and other properties of the ferredoxin from C. tartarivorum and C. thermosaccharolyticum were previously reported (1). Two striking features pertinent to the present study are the heat stability and the occurrence of histidine for the first time in the bacterial ferredoxins. It was suggested that heat stability might arise from a more stable chelate structure than is found in mesophilic ferredoxin molecules.
The present study has revealed that with the exception of several point mutations the over-all sequence of the thermophilic C. tartarivorum ferredoxin compares well with the sequence of the mesophilic ferredoxins.
No gross changes in the primary structure have accompanied the acquisition of thermostability by the C. tartarivorum ferredoxin molecule. The cysteinyl residues which anchor the chelate structure can be superimposed upon the cysteinyl residues of the bacterial ferredoxins sequenced to date (4,5,6,18) (see Fig. 6).
We have looked for amino acid replacements which might account for the heat stability especially around the cysteinyl residues which are involved in the chelation of iron.
Two such replacements were found : serine which usually precedes cysteine* was replaced by glutamic acid and alanine which usually occurs after cysteine" was replaced by glutamine. Other drastic changes are also shown in Fig. 6. It should be noted that the basic residues are not located in the vicinity of the cysteinyl residues in the primary structure.
The two histidines are found at posit,ions 2 and 24 and the two lysines are found at positions 29 and 53. Summarized in Table X are the types of residues found in the various ferredoxins.
The classification of residues is based on definitions adopted by Dickerson and Geis (19). It should be noted that again the C. tartarivorum ferredoxin has the greatest content of basic amino acid residues of any ferredoxin sequenced to date and therefore has the least negative net charge.
Admittedly   In the hope that a more meaningful comparison between thermophilic ferredoxins and the mesophilic ferredoxins can be made, sequence studies on the thermostable ferredoxin from C. thermosaccharolyticum have been initiated. Let us now examine the similarities in the sequences of the various bacterial ferredoxins available.
The results, expressed as the number of residues shared in common, are presented in Table XI.
The procedure for aligning the various ferredoxins has been discussed in a previous report from our laboratory (4). A possible phylogenetic relationship based upon the ferredoxins is shown in Fig. 7 further tested by structural studies on other sets of proteins, the rubredoxins and flavodoxins, which are also being sequenced.