Metal Ion Binding to a Zinc Finger Peptide Containing the Cys-X2-Cys-X4-His-X4-Cys Domain of a Nucleic Acid Binding Protein Encoded by theDrosophilaFw-Element

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.

242, 385–389 (1998)

RC977974

Metal Ion Binding to a Zinc Finger Peptide Containing the Cys-X2-Cys-X4-His-X4-Cys Domain of a Nucleic Acid Binding Protein Encoded by the Drosophila Fw-Element A. Bavoso,* A. Ostuni,* G. Battistuzzi,† L. Menabue,† M. Saladini,† and M. Sola†,1 *Department of Chemistry, University of Basilicata, Via N. Sauro, 85, 85100 Potenza, Italy; and †Department of Chemistry, University of Modena, Via Campi 183, 41100 Modena, Italy

Received November 17, 1997

The metal binding properties of a 18-residue zinc finger peptide containing a CCHC box which reproduces one of the cysteine-rich domains of a putative nucleic acid binding protein encoded by the Fw transposable element from Drosophila melanogaster were investigated through electronic and 1H NMR spectroscopy. Dissociation constants of 2({1)110012 M and 4({1)11007 M were determined for the Zn2/ and Co2/ adduct, respectively. These values are similar to those for other CCHC-peptides investigated previously, although the length of the spacer between the second cysteine and the histidine apparently exerts some influence on the spectral properties and on the stability of the Co2/-peptide adduct. The 1H NMR spectrum of the present Co2/-derivative contains a number of well resolved hyperfine-shifted resonances between 350 and 050 ppm which arise from the metal binding residues and nearby groups. These peaks can in principle be profitably exploited to monitor protein-nucleic acid interactions. q 1998 Academic Press

A wide variety of nucleic acid-binding and gene regulatory proteins contains a number of small domains structurally organized around tetrahedral Zn(II) ion(s) coordinated by Cys and His residues, termed ‘‘zinc finger’’ domains (1-3). An increasingly large number of proteins of this class binding to double-stranded nucleic acids are being recognized. The most extensively characterized species are eukaryotic transcription factors containing repeated mononuclear Zn domains with a CCHH binding set (C Å cysteine, H Å histidine), also termed ‘‘classical’’ zinc finger domains. However, other zinc binding motifs, such as CCCC centers and Zn2(Cys)6 clusters are known to be involved in protein-nucleic acid 1 Corresponding author. Fax: //39-59-373-543. E-mail: sola@ unimo.it.

recognition (1-3). Nucleocapsid proteins of retroviruses are basic species of low-molecular weight (from about 6 to 9 kDa) containing one or two copies of the zinc-binding motif Cys-X2-Cys-X4-His-X4-Cys (called ‘‘CCHC box’’) which, at variance with above, bind to single stranded nucleic acids (4-11). Proteins containing the same binding motif are also encoded by Drosophila melanogaster transposable elements (4, 12-16). A number of investigations have focused on the binding properties of synthetic peptides containing zinc finger sequence(s) toward spectroscopically active metals such as Co2/, Ni2/, Fe2/ and Cd2/ and on the structural features of the metal adducts (4, 17-23). The metal binding affinities of peptides containing the CCHH and CCHC box were found to differ to some extent (17). Moreover, the peptide domain containing the CCHC box was shown to fold into a globular structure (4, 7, 24-26) which differs significantly from the antiparallel b hairpin followed by a helix typical of ‘‘classical’’ CCHH zinc finger species (27). Here, we report on the metal binding properties of a 18-residue finger peptide containing a CCHC box which reproduces one of the cysteine-rich domains of a putative nucleic acid binding protein encoded by a 366-bp open reading frame present at the truncated 5* terminus of the Drosophila Fw element (12). This peptide, ValGlnCysThrAsnCysGlnGluTyrGlyHisThrArgSerTyrCysThrLeu (DF hereafter), has X spacers scarcely sequence related with those of the peptides from retroviral species studied previously (4, 24, 25) and a lower isoelectric point (pI Å 6.7 vs. 7.9-9.7) (4). We determined the binding affinity of DF toward Zn2/ and Co2/ through electronic spectroscopy and compared it with that of other zinc finger peptides. Moreover, we have detected the hyperfine-shifted 1H NMR resonances arising from the metal-binding and contiguous residues for the Co2/-DF adduct. Location of these signals, which are highly informative of the structural and electronic properties of the metal site (28, 29), is unprecedented for this class of metal centers.

385

0006-291X/98 $25.00 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.

AID

BBRC 7974

/

6945$$$461

12-21-97 09:16:14

bbrcg

AP: BBRC

Vol. 242, No. 2, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 1. Absorption spectra of 211004 M DF in 0.1 M phosphate buffer, pH 8, titrated with Co2/ dichloride. Titration was performed under strictly anaerobic conditions. Spectra were corrected for the absorption due to the free peptide. T Å 25 7C.

MATERIALS AND METHODS Reagents. N-9-fluorenylmethoxycarbonyl (Fmoc) aminoacids and 4-hydroxymethylphenoxy-methyl-copoly(styrene-1% divinylbenzene) (HMP) resin were purchased from Nova Biochem (Switzerland). Nmethylpyrrolidone (NMP), dicycloexylcarbodiimide (DCC), 4-dimethylaminopyridine (DMAP), 1-hydroxybenzotriazole (HOBt), 1,2ethanedithiol (EDT), dithiothreitol (DTT) and trifluoroacetic acid (TFA) were from Applied Biosystem. Peptide synthesis and purification. The peptide was synthesized with the stepwise solid method on a Applied Biosystems 431A peptide synthesizer using Fmoc chemistry. The C-terminal aminoacid was attached to the HMP resin with DCC in the presence of DMAP as catalyst. The subsequent Fmoc aminoacids were bound using DCC/ HOBt chemistry. Deprotection of the Fmoc group was obtained with 20% (w/w) piperidine in NMP. The peptide was cleaved from the resin upon treatment with TFA (88%), phenol (4%), EDT (2%), thioanisole (3%) and H2O (3%) for three hour at ambient temperature; this procedure also allows detachment of the protecting groups of the side chains. The mixture was then filtered and the solution was concentrated under vacuum. The peptide was then precipitated with cold ether, suspended in 0.1% TFA and lyophilized. The peptide, dissolved in Tris-HCl buffer in the presence of 50 equivalents of DTT, was purified on an analytical and semipreparative scale by reversephase high performance liquid chromatography on a Beckman Model 110B apparatus with a Bondclone C18 column 3.9 1 300 mm and on a Vydac C18 column 10 1 250 mm, respectively. The mobile phase consisted of solution A (0.1% TFA in H2O) and solution B (0.05%

TFA in CH3CN). The pure peptide was lyophilized under vacuum and stored in the presence of 5 equivalents of solid DTT under anaerobic conditions. The peptide was sequenced on an Applied Biosystem 491A instrument. A molecular mass of 2106.8 Da, consistent with the theoretical value of 2106.3 Da was obtained by electrospray mass spectrometry. Metal binding studies. All peptide manipulations described below were performed under an atmosphere of 85% nitrogen, 10% hydrogen and 5% CO2 in an anaerobic chamber (Plas-Lab, Lansing, MI, USA) to avoid peptide oxidation. DTT was separated from the reduced peptide through gel-filtration chromatography on a Sephadex G-15 column (5180 mm) in 0.1 M phosphate buffer, pH 8. The reduced peptide showed an extinction coefficient of 2450 M01 cm01 at 280 nm. The dissociation constant for the Co2/ adduct was determined by titrating the reduced peptide in 0.1 M phosphate buffer at pH 8 with the metal ion and monitoring the absorption spectrum. This pH was chosen in order to increase the solubility of the peptide, which is almost uncharged at pH 7. At higher pH values, the formation of the metal-peptide adduct suffers competition with metal hydroxide precipitation. The constant for Zn2/ binding was obtained by following the bleaching of the absorption spectrum of the Co2/-adduct upon addition of the metal, as reported elsewhere (21) The electronic spectra were recorded on a Perkin-Elmer Lambda 9 spectrophotometer. NMR spectra. NMR samples of the Co2/-DF adduct were prepared by adding one equivalent of cobalt dichloride to 500 mL of the reduced peptide (about 0.1 mM) in 0.1 M phosphate buffer at pH 8. 1 H NMR measurements were carried out on Bruker AMX-400 and DPX-200 spectrometers at 400.13 and 200.03 MHz, respectively. Typ-

386

AID

BBRC 7974

/

6945$$$462

12-21-97 09:16:14

bbrcg

AP: BBRC

Vol. 242, No. 2, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

ical acquisition parameters were as follows: spectral width, 125 kHz; pulse width, 6 ms (907 pulse); pulse delay, 0.2-0.4 s. Spectra were referenced to tetramethylsilane (TMS) after calibration against the water peak, set at 4.78 ppm from TMS (at 20 7C). Suppression of the water peak was achieved using the super-WEFT pulse sequence (30). Longitudinal relaxation times were determined with the inversion recovery sequence preceded by a presaturation pulse.

RESULTS AND DISCUSSION Addition of Co2/ to the reduced peptide in anaerobic conditions at pH 8 originates absorption bands at 310 (e Å 3100 M01 cm01), 355 (shoulder), 612 (shoulder), 645 (e Å 460 M01 cm01) and 697 (e Å 380 M01 cm01) nm (Fig. 1). The first two bands at higher energy can be attributed to cysteinate to metal charge transfer transitions, while the three components of the spectral envelope between 600 and 700 nm can be assigned to 4 A2 to 4T1(P) ligand field transitions typical of synthetic tetrahedral Co2/ complexes with thiolate and imidazole ligands (4, 31, 32). The spectrum is almost identical to those of the Co2/ adducts of the nucleocapsid protein p10 from Rauscher murine leukemia virus (RaMLV) (33) and of the 18 aminoacid fragment of the same protein containing the CCHC binding motif CysX2-Cys-X4-His-X4-Cys (4). Titration of reduced DF with Co2/ was followed spectrophotometrically (Fig. 2A). Band intensities increase linearly with increasing metal concentration and level off at the stoichiometric Co2//peptide ratio of 1:1. The data could be fitted by non linear least-squares methods which yielded a dissociation constant of KdCo Å 4({1)11007 M. This value compares reasonably well with that determined for the RaMLV peptide (KdCo Å 111006 M) (4) (at least part of the difference may arise from the fact that the two constants refers to different conditions of ionic composition and pH). Instead, the Co2/adduct of the His24 to Cys variant of CP-1 [CP1(CCHC)] (where CP-1 is is the consensus zinc finger peptide ProTyrLysCys4ProGluCys7GlyLysSerPheSerGluLysSerAspLeuValLysHis 20GlnArgThrHis24 ThrGly prototypic of the zinc binding domains of eukaryotic transcription factors (20)) has a lower dissociation constant (KdCo Å 6.311008 M) (17). It may be proposed that the lower affinity of the peptides of retroviral origin for Co2/ is due to some constraints to the coordination geometry imposed by the only four-residue spacing between the second Cys and the His ligands instead of the 12-residue spacing in CP-1(CCHC). Moreover, the three d-d bands in the electronic spectra of the Co2/-CP-1(CCHC) adduct are spread over a wider region (from 580 to 724 nm) and are better resolved than those of the Co2/ adducts of the RaMLV and DF peptides (17, 20, 22). Thus, it appears that the length of the spacer between the second cysteine and the histidine in the sequence of the CCHC domain exerts some influence on the ligand field transitions of the metal chromophore and on the metal affinity

FIG. 2. Plot of the absorbance at 645 nm for (A) 211004 M DF in 0.1 M phosphate buffer, pH 8, in the presence of added Co2/ (see Fig. 1), as a function of the [Co2/]/[DF] ratio, and (B) 211004 M Co2/DF in 0.1 M phosphate buffer, pH 8, in the presence of added Zn2/ as a function of the [Zn2/]/[DF] ratio. T Å 25 7C.

of the peptide, while residue substitutions within the conserved retroviral sequence motif Cys-X2-Cys-X4His-X4-Cys have little effect. The affinity of DF for Zn2/ was determined by titrating the Co2/ adduct with the above metal in the presence of a known excess of Co2/ ion (50 fold), and following the disappearance of the absorption bands due to the displacement of Co2/ by Zn2/, which is known to bind more tightly to these peptides (17, 20, 21) (Fig. 2B). The Zn2/ ion in fact does not undergo loss in ligand field stabilization energy on passing from the aquoion to a tetrahedral coordination, as does Co2/, as pointed out elsewhere (17). The dissociation constant was determined to be: KdZn Å 2({1)110012 M, which is close to that determined for CP-1(CCHC) (KdZn Å 3.2({1)110012 M) (17). The paramagnetic Co2/ ion is profitably exploited as NMR probe for the investigation of metal sites in metalloproteins, due to the contact and pseudocontact contributions to the chemical shift and nuclear relaxation of

387

AID

BBRC 7974

/

6945$$$462

12-21-97 09:16:14

bbrcg

AP: BBRC

Vol. 242, No. 2, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 3. 1H NMR spectrum of 211004 M Co2/-DF in 0.1 M phosphate buffer, pH 8. (A) 200 MHz spectrum showing the details of the high- and low-frequency paramagnetic region. Number of scans, 16k. (B) 400 MHz spectrum showing the hyperfine-shifted signals close to the diamagnetic region. Number of scans, 8k. T1 values (ms) are indicated below each peak. Shaded signals are those which disappear in the spectrum recorded in D2O, and correspond to exchangeable amide protons. T Å 27 7C.

the protons in the surroundings of the metal, which provide valuable structural information (28, 29). The 200 MHz 1H NMR spectrum of the Co2/-DF adduct contains a number of strongly hyperfine-shifted resonances between 350 and 050 ppm (Fig. 3A). T1 vales are submillisecond (from below 0.1 to 0.3 ms), except peak h (T1 Å 15 ms). Other well resolved and more slowly relaxing resonances are located from 10 to 25 and from 0 to 015 ppm (Fig. 3B). Similar spectral patterns were observed for a Co2/-substituted rubredoxin which possesses a tetrahedral Co-S(Cys)4 site (29, 34), and for Co2/-metallothioneins which contain polynuclear Co-thiolate clusters in which each metal is again tetrahedrally coordinated by four cysteine residues (35, 36). By reference to the

above systems, six of the eight nonexchangeable signals which fall above 40 ppm (a-g, i) should correspond to the b-CH2 groups of the three cysteine ligands. The remaining two peaks reasonably arise from a-CH proton(s) of binding cysteine(s) and/or ring protons of the histidine ligand which likely binds through the Ne2 atom (24). The exchangeable peak h at 63 ppm can be attributed to the Nd1H proton of the above histidine. This 1H NMR spectrum of a Co2/-substituted zinc finger peptide containing the complete set of hyperfine-shifted resonances arising from the metal binding residues is unprecedented. Previous 1H NMR work on the Co2/CP-1 adduct focused only on the resonances affected by pseudocontact shift, allowing determination of the orien-

388

AID

BBRC 7974

/

6945$$$463

12-21-97 09:16:14

bbrcg

AP: BBRC

Vol. 242, No. 2, 1998

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

tation and anisotropy of the magnetic susceptibility tensor (18). Assignment of the hyperfine-shifted resonances of the Co2/-DF adduct could not be achieved due to several unfavorable factors, namely: i) the small stead state NOEs (most likely below detection) which are expected to correlate these signals, due to the remarkably low T1 values and the low molecular weight of the adduct (28, 29); ii) the scarce solubility of DF in aqueous solution at neutral or slightly alkaline pH (max 0.3 mM at pH 8); iii) the instability of the Co2/-DF derivative which, in inert atmosphere, decomposes appreciably already after 6-8 hours after metal addition. Nevertheless, the present work shows that, despite tetrahedral four-coordination of the high-spin Co2/ ion induces a larger broadening of the paramagnetic resonances as compared to five or six-coordination (28, 29), a number of well resolved 1H NMR peaks due to the protons of the metal site experiencing contact and pseudocontact shift can be detected for these adducts. Similar spectral features should also characterize the Co2/-derivatives of the whole proteins for which measurable NOEs should in principle be observed especially at high magnetic fields. Thus, paramagnetic 1H NMR can in principle be successfully exploited for obtaining structural information in solution for Co2/-substituted zinc finger proteins and for the investigation of their interaction with nucleic acids. ACKNOWLEDGMENTS This study was supported by the University of Modena (Finanziamento Ricerca Avanzata) and by the National Research Council of Italy (Progetto Strategico Tecnologie Chimiche Innovative).

REFERENCES 1. 2. 3. 4. 5.

6.

7. 8.

Berg, J. M., and Shi, Y. (1996) Science 271, 1081–1085. Berg, J. M. (1995) Acc. Chem. Res. 28, 14–19. Klug, A., and Schwabe, J. W. R. (1995) FASEB J. 9, 597–604. Green, L. M., and Berg, J. M. (1989) Proc. Natl. Acad. Sci. USA 86, 4047–4051, and references therein. Darlix, J.-L., Gabus, C., Nugeyre, M.-T., Clavel, F., and Barre´Sinoussi, F. (1990) J. Mol. Biol. 216, 689–699 and references therein. De Rocquigny, H., Ficheux, D., Gabus, C., Allain, B., FournieZaluski, M. C., Darlix, J.-L., and Roques, B. P. (1993) Nucleic Acids Research 21, 823–829 and references therein. Morellet, N., Jullian, N., De Rocquigny, H., Maigret, B., Darlix, J. L., and Roques, B. P. (1992) EMBO J. 11, 3059–3065. Summers, M. F. (1991) J. Cell. Biochem. 45, 41–48.

9. Karpel, R. L., Henderson, L. E., and Oroszlan, S. (1987) J. Biol. Chem. 262, 4961–4967. 10. Prats, A. C., Housset, V., De Billy, G., Cornille, F., Prats, H., Roques, B. P., and Darlix, J.-L. (1991) Nucleic Acids Res. 19, 3533–3541. 11. Prats, A. C., Sarih, L., Gabus, C., Litvak, S., Keith, G., and Darlix, J. L. (1988) EMBO J. 7, 1777–1783. 12. Di Nocera, P. P., and Casari, G. (1987) Proc. Natl. Acad. Sci. USA 84, 5843–5847. 13. Di Nocera, P. P. (1988) Nucleic Acids Res. 16, 4041–4053. 14. Fawcett, T., Lister, C. K., Kellett, E., and Finnegan, D. J. (1986) Cell 47, 1007–1015. 15. Mount, S. M., and Rubin, G. M. (1985) Mol. Cell. Biol. 5, 1630– 1638. 16. Rubin, G. M. (1983) in Mobile Genetic Elements (Shapiro, J. A., Ed.) pp. 329–361. Academic Press, New York. 17. Krizek, B. A., Merkle, D. L., and Berg, J. M. (1993) Inorg. Chem. 32, 937–940. 18. Harper, L. V., Amann, B. T., Kilfoil, V. J., and Berg, J. M. (1993) J. Am. Chem. Soc. 115, 2577–2580. 19. Krizek, B. A., and Berg, J. M. (1992) Inorg. Chem. 31, 2984– 2986. 20. Krizek, B. A., Amann, B. T., Kilfoil, V. J., Merkle, D. L., and Berg, J. M. (1991) J. Am. Chem. Soc. 113, 4518–4523. 21. Berg, J. M., and Merkle, D. L. (1989) J. Am. Chem. Soc. 111, 3759–3761. 22. Blasie, C. A., and Berg, J. M. (1997) Biochemistry 36, 6218– 6222. 23. South, T. L., Kim, B., and Summers, M. F. (1989) J. Am. Chem. Soc. 111, 395–396. 24. Summers, M. F., South, T. L., Kim, B., and Hare, D. R. (1990) Biochemistry 29, 329–340. 25. South, T. L., Blake, P. R., Hare, D. R., and Summers, M. F. (1991) Biochemistry 30, 6342–6349. 26. Omichinski, J. G., Clore, G. M., Sakaguchi, K., Appella, E., and Gronenborn, A. M. (1991) FEBS Lett. 292, 25–30. 27. Kim, C. A., and Berg, J. M. (1996) Nat. Struct. Biol. 3, 940–945. 28. Bertini, I., and Luchinat, C. (1996) Coord. Chem. Rev. 150, 1– 296. 29. Bertini, I., Turano, P., and Vila, A. J. (1993) Chem. Rev. 93, 2833–2932. 30. Inubushi, T., and Becker, E. D. (1983) J. Magn. Reson. 51, 128– 133. 31. Bertini, I., and Luchinat, C. (1984) Adv. Inorg. Biochem. 6, 71– 111. 32. Corwin, D. T., Jr., Gruff, E. S., and Koch, S. A. (1988) Inorg. Chim. Acta 155, 5–6. 33. Roberts, W. J., Pan, T., Elliot, J. I., Coleman, J. E., and Williams, K. R. (1989) Biochemistry 28, 10043–10047. 34. Moura, I., Teixeira, M., LeGall, J., and Moura, J. J. G. (1991) J. Inorg. Biochem. 44, 127–139. 35. Bertini, I., Luchinat, C., Messori, L., and Vasak, M. (1989) J. Am. Chem. Soc. 111, 7296–7300. 36. Bertini, I., Luchinat, C., Messori, L., and Vasak, M. (1993) Eur. J. Biochem. 211, 235–240.

389

AID

BBRC 7974

/

6945$$$463

12-21-97 09:16:14

bbrcg

AP: BBRC