conversion of desulforedoxin into a rubredoxin center

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS 231, 679–682 (1997)ARTICLE NO. RC976171

Conversion of Desulforedoxin into a Rubredoxin Center

Lian Yu,* Matthew Kennedy,* Christopher Czaja,* Pedro Tavares,† Jose J. G. Moura,†Isabel Moura,† and Frank Rusnak*,1

*Section of Hematology Research and the Department of Biochemistry and Molecular Biology, Mayo Clinic andFoundation, Rochester, Minnesota 55905; and †Departmento de Quımica and Centro de Quımica Fina e Biotecnologia,Faculdade de Ciencias e Tecnologia, Universidade Nova de Lisboa, 2825 Monte de Caparica, Portugal

Received December 30, 1996

crystal structure of Dx has found that the Sg-Fe-SgRubredoxin and desulforedoxin both contain an bond angle involving C28 and C29 is unusually large

Fe(S-Cys)4 center. However the spectroscopic proper- at 121.57 (6). Although the x-ray structure shows a dis-ties of the center in desulforedoxin differ from ru- tortion away from tetrahedral geometry, it is notbredoxin. These differences arise from a distortion of known whether it results from the lack of interveningthe metal site hypothesized to result from adjacent residues. In order to address the above hypothesis, wecysteine residues in the primary sequence of desulfo-

have constructed mutants of Dx in which one or tworedoxin. Two desulforedoxin mutants were generatedresidues have been inserted between the C-terminusin which either a G or P-V were inserted between adja-cysteines. Dx mutants exhibit uv/visible and EPR spec-cent cysteines. Both mutants exhibited optical spectratroscopic features distinct from wild type and virtuallywith maxima at 278, 345, 380, 480, and 560 nm while theidentical to Rd.low temperature X-band EPR spectra indicated high-

spin Fe3/ ions with large rhombic distortions (E/D Å0.21-0.23). These spectroscopic properties are distinct MATERIALS AND METHODSfrom wild type desulforedoxin and virtually identicalto rubredoxin. q 1997 Academic Press

Methods. DNA sequencing used the Sequenase 2.0 DNA kit(USB, Cleveland, OH) and [35S]-deoxyadenosine 5*-(a-thio)triphos-phate (500 Ci/mmol; NEN, Wilmington, DE). EPR spectra were re-corded on a Bruker ESP300E spectrometer equipped with an Oxford

The prototype Fe(S-Cys)4 center is rubredoxin (Rd), Instruments ESR 900 continuous flow cryostat. Iron analyses werea polypeptide which binds a single iron atom in approx- determined by the metals laboratory at the Mayo Clinic using induc-

tively coupled plasma emission spectroscopy (7). Protein concentra-imate tetrahedral geometry (1). Another protein con-tion was determined by amino acid analysis.taining the Fe(S-Cys)4 center is Dx (Dx) from Desulfovi-

bro gigas, an a2 homodimer consisting of two, 36 resi- Site directed mutagenesis of Dx. Mutagenesis was carried outaccording to Higuchi (9) to insert either a G (mutant 1), or P-Vdue polypeptide chains, each which provides four(mutant 3) between C28 and C29. These oligonucleotides (mutantcysteines for binding a single iron atom (2-8). Signifi-codons underlined) were used for mutant 1; 5*-GGCACGCTGGTG-cant differences exist in the optical and EPR spectra TGCGGCTGCGGCGAGGATATG-3*, and 5*-CATATCCTCGCCGCA-

of the Fe3/ complexes of Dx and Rd (2,3,5,7) due to GCCGCACACCAGCGTGCC-3*. For mutant 3, 5*-GGCACGCTG-differences in the metal binding motif. In Rd the cys- GTGTGCCCCGTGTGCGGCGAGGATATG-3*, and 5*-CATATCCTC-

GCCGCACACGGGGCACACCAGCGTGCC-3* were used. The dsrteine ligands occur in pairs at the N- and C-terminii,gene cloned into pUC19 (7) was used as template and M13/pUCeach with the sequence C-X-X-C. In Dx the cysteineprimers (New England Biolabs, Inc., Beverly, MA) were used aspair at the N-terminus shares this pattern but the pair flanking primers. Following amplification the mutant DNA frag-

at the C-terminus consists of adjacent cysteines (C28 ments were digested with Xba I and Pst I and subcloned into pUC19.Recombinant clones were identified and the mutant gene from aand C29), leading to the hypothesis that the absencepositive clone excised using Nde I and Hind III and subcloned intoof intervening residues between the C-terminus pairpT7-7 (10) to give the recombinant vectors DSRM1ET77 (mutant 1)causes a distortion of the metal center. In fact, theand DSRM3D4T77 (mutant 3). These plasmids were transformedinto BL21 (DE3) cells (Novagen Inc., Madison, WI) (11) and mutantDx proteins purified as described (7).1 To whom correspondence should be addressed. Fax: (507) 284-

8286. Mass spectrometry of Dx mutants. MALDI-TOF mass spectrome-try was carried out in the positive ion mode on a Bruker BiflexAbbreviations used: Dx, desulforedoxin; MALDI-TOF, matrix as-

sisted laser desorption ionization-time of flight mass spectrometry; mass spectrometer in the Mayo Clinic Biomedical Mass SpectrometryFacility. Typically 1 mL of a 50 mM protein solution was mixed withMr, molecular mass; Rd, rubredoxin; TCA, trichloroacetic acid.

0006-291X/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

679

AID BBRC 6171 / 691c$$$781 02-07-97 00:49:11 bbrcg AP: BBRC

Vol. 231, No. 3, 1997 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

486, and 560 nm and a shoulder at 345 nm. The extinc-tion coefficients shown in figure 2 were calculatedbased on the iron content of each sample (typically 0.6 -0.7 mol Fe/mol protein) from at least two separatelyreconstituted samples.

EPR of Dx mutants. EPR spectra of Fe3/-Dx wildtype, mutant 1, and mutant 3 are shown in figure 3.The spectrum of mutant 3 was remarkably similar to

FIG. 1. MALDI-MS of Dx mutant 1 (A) and Dx mutant 3 (B).Samples were prepared as described in the Materials and Methodssection. The measured Mr for each sample is indicated by an arrowand the predicted apoprotein Mr (not including the N-terminus me-thionine which is removed during expression in E. coli ) is shown inthe box. Mr, molecular mass.

4 mL of a-cyanohydroxycinnamic acid in acetonitrile/water/trifluoro-acetic acid (70/30/0.1, v/v/v) and 3 mL spotted onto the target.

Reconstitution with iron. Each mutant protein was reconstitutedwith iron by TCA precipitation followed by incubation with 0.05Mb-ME and É2 equivalents of FeSO4 as described (5,12).

RESULTS

Mutagenesis of Dx. Two mutants of Dx were gener-ated in which the spacing between C28 and C29 wasaltered. Mutant 1 represents the simplest change withthe insertion of a G and has the sequence (C-G-C)28-30.The alteration for mutant 3 corresponds to the C-termi-nus cysteine pair of D. gigas Rd (1) with the sequence(C-P-V-C)28-31 . Following mutagenesis, the resultingDNA was cloned into the expression vector pT7-7, ex-pressed in E. coli BL21(DE3) cells, and each mutantprotein purified to homogeneity. Both mutant proteinswere expressed exclusively as the Zn2/ form. Massspectrometry indicated a homogenous preparation ofpolypeptide exhibiting the correct mass (Figure 1).Mass spectrometry of mutant 1 yielded a molecularion with Mr of 3861.7 (calculated Mr Å 3861.4), whilemutant 3 afforded a molecular ion with Mr Å 3999.7(calculated Mr Å 4000.6).

UV/visiblespectraofDxmutants1and3. Reconstitu-tion with iron yielded the Fe3/ complexes for both mu-tants 1 and 3. The optical spectra of both mutants (fig-ure 2) were distinct from the spectrum of wild type Dx(2,5,7) yet resembled the spectrum of Rd (13-15). The

FIG. 2. Optical spectra from 250 to 700 nm of the Fe3/ complexesFe3/ complex of Dx mutant 1 exhibited absorbanceof wild type Dx (DxWT), Dx mutant 1 (DxM1), Dx mutant 3 (DxM3),maxima at 278, 345, 484, and 560 nm with a shoulder and D. gigas rubredoxin (Rd). Extinction coefficients are expressed

also evident at 380 nm. The optical spectrum of mutant relative to the iron content of the sample and represent the averageof at least two determinations.3 was similar in appearance with maxima at 278, 375,

680



the EPR spectra of Rd from P. oleovorans (16) and M.aerogenes (14) and is attributed to the Fe(S-Cys)4 site.The spectrum of mutant 3, with resonances at g Å9.28, 4.94, 4.31, and 3.5, is characteristic of a high-spin (S Å 5/2) ferric ion with a rhombicity (E/D) of0.23. The observation that the intensity of the low fieldresonance was inversely proportional to temperaturefrom 3.6 to 12 K indicates that it arises from transi-tions between a ground state Kramers doublet. Thusthe sign of D is positive. The EPR spectrum of mutant1, with resonances at g Å 9.21 and 4.3, is attributedto the Fe(S-Cys)4 site by comparison with the EPRspectra of mutant 3 and Rd. Based on the g-value andtemperature dependence, the g Å 9.21 resonance alsoarises from the {1/2 ground doublet of a high-spinferric center with Dú0 and E/D É 0.21. The shapeand intensity of this resonance, especially the sharpderivative feature at g Å 4.3, resembles the EPR spec-tra of Rd from C. ethylica (17) and the Rd-like centerof rubrerythrin (18).

DISCUSSION

Dx from D. gigas contains a single iron atom coordi-nated by four cysteinyl sulfur atoms (2,4-7). As such,it resembles the iron-sulfur protein Rd which also con-tains a single iron atom coordinated by four sulfuratoms in a tetrahedral geometry (1). However, the spec-troscopic properties of Dx are distinct from Rd. Al-though the optical spectra of both Dx (2,5,7) and Rd(13-15) are characterized by ligand-to-metal chargetransfer transitions in the visible region of the spec-trum, there are differences in both the intensities and FIG. 3. EPR spectra of the Fe3/ complexes of wild type Dx (DxWT),

mutant 1 (DxM1), and mutant 3 (DxM3) with g-values of selectedpositions of absorbance maxima. The low temperatureresonances indicated. The buffer was 50 mM TrisCl, 50 mM NaCl,X-band EPR spectra of Dx are characterized by reso-pH 7.5 for all samples. Spectrometer conditions: temperature, 9 K fornances with g-values at 7.6, 5.7, 4.1, and 1.8, and are wild type and 3.6 K for both mutants; microwave frequency, 9.433

attributed to a high-spin Fe3/ ion with a rhombicity GHz; microwave power, 2.0 mW for wild type and 8.0 mW for bothmutants; modulation frequency and amplitude, 100 kHz at 10 G.parameter (E/D) of 0.08 (3,5,7). Rd, on the other hand,

exhibits resonances at g Å 9.4, 4.7, 4.3, and 4.0, alsodue to a high-spin ferric ion but with a much largerrhombicity (14,16-18). geometry (6). Although the S-Fe bond distances are

essentially the same as in Rd, the bond angle betweenAn explanation for the spectroscopic differences be-tween Rd and Dx was originally hypothesized to result C28 and C29 is 121.57, significantly widened when com-

pared to Rd.from geometric strain at the Dx metal center arisingfrom a difference in the metal binding motif (5). In Rd, The two Dx mutants generated in this study support

the hypothesis that the lack of intervening residuesthis motif consists of two pair of cysteine residues, eachseparated by two residues at the N (CTVCG)6-10 and C between C28 and C29 is responsible for imparting

strain at the metal and the subsequent manifestation(CPVCG)39-43 terminii (1). The metal ion in Rd residesin an approximate tetrahedral geometry with Sg-Fe- of the unique optical and EPR properties of Dx. The

insertion of either G (mutant 1) or P-V (mutant 3) be-Sg bond angles ranging from 103.4 to 114.5 degreesand an average S-Fe bond distance of 2.29 A { 0.022 tween adjacent cysteines resulted in polypeptides that

still retained their ability to complex iron, but whichA. In Dx, the spacing between cysteine residues of theN terminus pair is preserved (CELCG)9-13 while the now exhibited spectroscopic characteristics that can

best be described as ‘‘Rd-like’’. Thus, the optical spectrapair at positions 28 and 29 consists of adjacent cys-teines (4,19). The recent determination of the crystal of both mutants exhibit an apparent resolution of the

370 nm absorption band observed in wild type Dxstructure of Dx at 1.8 A resolution has confirmed thatthe metal ion in Dx exists in a distorted tetrahedral (2,5,7), into two overlapping peaks at ca. 345 and 380

681



5. Moura, I., Huynh, B. H., Hausinger, R. P., Le Gall, J., Xavier,nm characteristic of Rd (figure 3). An increased separa-A. V., and Munck, E. (1980) J. Biol. Chem. 255, 2493–2498.tion between absorption bands in the 480-600 nm re-

6. Archer, M., Huber, R., Tavares, P., Moura, I., Moura, J. J. G.,gion is also apparent in the mutant versus wild typeCarrondo, M. A., Sieker, L. C., LeGall, J., and Romao, M. J.

spectra. The EPR spectra of both mutant proteins are (1995) J. Mol. Biol. 251, 690–702.characteristic of high-spin ferric ions with E/D É0.22, 7. Czaja, C., Litwiller, R., Tomlinson, A. J., Naylor, S., Tavares, P.,a value similar to that obtained for Rd by EPR (16) or LeGall, J., Moura, J. J. G., Moura, I., and Rusnak, F. (1995) J.

Biol. Chem. 270, 20273–20277.Mossbauer (20) analyses, and significantly larger than8. Tavares, P., Wunderlich, J. K., Lloyd, S. G., LeGall, J., Moura,the value observed for wild type Dx.

J. J. G., and Moura, I. (1995) Biochem. Biophys. Res. Comm. 208,These results confirm the hypothesis that the lack of680–687.intervening residues between C-terminal cysteines in

9. Higuchi, R., Krummel, B., and Saiki, R. K. (1988) Nuc. Acid Res.Dx are responsible for the differences in uv/vis and 16, 7351–7367.EPR spectra compared to Rd. An NMR solution or x-ray 10. Tabor, S. (1990) in Current Protocols in Molecular Biology (Ausu-crystal structure of Dx mutants would be particularly bel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman,

J. G., Smith, J. A., and Struhl, K., Eds.), pp. 16.2.1–16.2.11,useful to determine metal-ligand bond lengths andGreene Publishing and Wiley Interscience, New York.angles for comparison with wild type Dx and Rd. These

11. Studier, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff,studies are currently in progress.J. W. (1990) Methods Enzymol. 185, 60–89.

12. Moura, I., Teixeira, M., Le Gall, J., and Moura, J. J. G. (1991)ACKNOWLEDGMENTS J. Inorganic Biochem. 44, 127–139.

13. Lovenberg, W., and Sobel, B. E. (1965) Proc. Natl. Acad. Sci.We wish to thank Larry Sieker for help in designing Dx mutants USA 54, 193–199.

and M. J. Romao for helpful discussions. We also thank M. Holmes 14. Bachmayer, H., Piette, L. H., Yasunobu, K. T., and Whiteley,and S. Naylor of the Mayo Biomedical Mass Spectrometry Facility H. R. (1967) Proc. Natl. Acad. Sci. USA 57, 122–127.for assistance with mass spectrometry. This work was supported by

15. Lovenberg, W., and Williams, W. M. (1969) Biochemistry 8, 141–the National Institutes of Health (GM46865) and the Junta Nacional148.de Investigacao CientıB fica e Tecnologica. F.R. was a PRAXIS fellow

16. Peisach, J., Blumberg, W. E., Lode, E. T., and Coon, M. J. (1971)at FCT-UNL.J. Biol. Chem. 246, 5877–5881.

17. Rao, K. K., Evans, M. C. W., Cammack, R., Hall, D. O., Thomp-REFERENCES son, C. L., Jackson, P. J., and Johnson, C. E. (1972) Biochem. J.

129, 1063–1070.1. Sieker, L. C. (1994) Methods Enzymol. 243, 203–216. 18. LeGall, J., Prickril, B. C., Moura, I., Xavier, A. V., Moura,2. Moura, I., Bruschi, M., Le Gall, J., Moura, J. J. G., and Xavier, J. J. G., and Huynh, B.-H. (1988) Biochemistry 27, 1636–1642.

A. V. (1977) Biochem. Biophys. Res. Comm. 75, 1037–1044. 19. Brumlik, M. J., Leroy, G., Bruschi, M., and Voordouw, G. (1990)3. Moura, I., Xavier, A. V., Cammack, R., Bruschi, M., and Le Gall, J. Bacteriol. 172, 7289–7292.

J. (1978) Biochim. Biophys. Acta 533, 156–162. 20. Debrunner, P. G., Munck, E., Que, L., and Schulz, C. E. (1977)in Iron–Sulfur Proteins (Lovenberg, W., Ed.), Vol. 3, pp. 381–4. Bruschi, M., Moura, I., Le Gall, J., Xavier, A. V., and Sieker,

L. C. (1979) Biochem. Biophys. Res. Comm. 90, 596–605. 417, Academic Press, New York.

682


conversion of desulforedoxin into a rubredoxin center

Documents